Journal of Cereal Science

Journal of Cereal Science
Volume 49, Issue 1, Pages 1-162 (January 2009)
1.
IFC - Editorial Board Page IFC
Rapid Communication 2. Degradation of cereal beta-glucan by ascorbic acid induced oxygen radicals Pages 1-3 Reetta Kivel, Fred Gates, Tuula Sontag-Strohm
Research Papers 3. Variation in Granule Bound Starch Synthase I (GBSSI) loci amongst Australian wild cereal relatives (Poaceae) Pages 4-11 F.M. Shapter, P. Eggler, L.S. Lee, R.J. Henry Effect of high temperature on albumin and globulin accumulation in the endosperm proteome of the developing wheat grain Pages 12-23 William J. Hurkman, William H. Vensel, Charlene K. Tanaka, Linda Whitehand, Susan B. Altenbach Accumulation of mixed linkage (1 3) (1 4)--d-glucan during grain filling in barley: A vibrational spectroscopy study Pages 24-31 Helene Fast Seefeldt, Andreas Blennow, Birthe Mller Jespersen, Bernd Wollenweber, Sren Balling Engelsen Mechanism of gas cell stabilization in bread making. I. The primary glutenstarch matrix Pages 32-40 Baninder S. Sroan, Scott R. Bean, Finlay MacRitchie Mechanism of gas cell stabilization in breadmaking. II. The secondary liquid lamellae Pages 41-46 Baninder S. Sroan, Finlay MacRitchie Expression of globulin-2, a member of the cupin superfamily of proteins with similarity to known food allergens, is increased under high temperature regimens during wheat grain development Pages 47-54 Susan B. Altenbach, Charlene K. Tanaka, William J. Hurkman, William H. Vensel Biochemical markers: Efficient tools for the assessment of wheat grain tissue proportions in milling fractions Pages 55-64 Youna Hemery, Valrie Lullien-Pellerin, Xavier Rouau, Jol Abecassis, Marie-Franoise Samson, Per man, Walter von Reding, Ccilia Spoerndli, Ccile Barron
4.
5.
6.
7.
8.
9.
10.
Lateral growth of a wheat dough disk under various growth conditions Pages 65-72 Amy Penner, Leaelaf Hailemariam, Martin Okos, Osvaldo Campanella Digestibility of protein and starch from sorghum (Sorghum bicolor) is linked to biochemical and structural features of grain endosperm Pages 73-82 Joshua H. Wong, Tsang Lau, Nick Cai, Jaswinder Singh, Jeffrey F. Pedersen, William H. Vensel, William J. Hurkman, Jeff D. Wilson, Peggy G. Lemaux, Bob B. Buchanan Fundamental study on protein changes taking place during malting of oats Pages 83-91 Christina Klose, Beatus D. Schehl, Elke K. Arendt Effect of milling, pasta making and cooking on minerals in durum wheat Pages 92-97 Francesco Cubadda, Federica Aureli, Andrea Raggi, Marina Carcea Starch granule size distribution of hard red winter and hard red spring wheat: Its effects on mixing and breadmaking quality Pages 98-105 Seok-Ho Park, Jeff D. Wilson, Bradford W. Seabourn Total phenolics, flavonoids, antioxidant capacity in rice grain and their relations to grain color, size and weight Pages 106-111 Yun Shen, Liang Jin, Peng Xiao, Yan Lu, Jinsong Bao Nucleotide polymorphisms in the waxy gene of NaN3-induced waxy rice mutants Pages 112-116 Toong Long Jeng, Chang Sheng Wang, Tung Hai Tseng, Min Tze Wu, Jih Min Sung Amaranth (Amaranthus hypochondriacus) as an alternative crop for sustainable food production: Phenolic acids and flavonoids with potential impact on its nutraceutical quality Pages 117-121 A.P. Barba de la Rosa, Inge S. Fomsgaard, Bente Laursen, Anne G. Mortensen, L. OlveraMartnez, C. Silva-Snchez, A. Mendoza-Herrera, J. Gonzlez-Castaeda, A. De Len-Rodrguez Relationship of milled grain percentages and flowering-related traits in rice Pages 122-127 Rodante E. Tabien, Stanley Omar P.B. Samonte, Emmanuel R. Tiongco Texture, processing and organoleptic properties of chickpea-fortified spaghetti with insights to the underlying mechanisms of traditional durum pasta quality Pages 128-133 Jennifer Ann Wood Impact of re-grinding on hydration properties and surface composition of wheat flour Pages 134-140 M. Mohamad Saad, C. Gaiani, J. Scher, B. Cuq, J.J. Ehrhardt, S. Desobry Identification of novel haze-active beer proteins by proteome analysis Pages 141-147 Takashi Iimure, Nami Nankaku, Megumi Watanabe-Sugimoto, Naohiko Hirota, Zhou Tiansu, Makoto Kihara, Katsuhiro Hayashi, Kazutoshi Ito, Kazuhiro Sato
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Mixing properties and dough functionality of transgenic lines of a commercial wheat cultivar expressing the 1Ax1, 1Dx5 and 1Dy10 HMW glutenin subunit genes Pages 148-156 Elena Len, Santiago Marn, Mara J. Gimnez, Fernando Piston, Marta Rodrguez-Quijano, Peter R. Shewry, Francisco Barro Wheat glutenin proteins assemble into a nanostructure with unusual structural features Pages 157-162 Sarah H. Mackintosh, Susie J. Meade, Jackie P. Healy, Kevin H. Sutton, Nigel G. Larsen, Adam M. Squires, Juliet A. Gerrard
23.

Aims and Scope
The Journal of Cereal Science was established in 1983 to provide an international forum for the publication of original research papers of high standing covering all aspects of cereal science related to the functional and nutritional quality of cereal grains and their products. The journal also publishes concise and critical review articles appraising the status and future directions of specific areas of cereal science and short rapid communications that present news of important advances in research. The journal aims at topicality and at providing comprehensive coverage of progress in the field. Research areas include: Composition and analysis of cereal grains in relation to quality in end use Morphology, biochemistry, and biophysics of cereal grains relevant to functional and nutritional characteristics Structure and physicochemical properties of functionally and nutritionally important components of cereal grains such as polysaccharides, proteins, oils, enzymes, vitamins, and minerals Storage of cereal grains and derivatives and effects on nutritional and functional quality Genetics, agronomy, and pathology of cereal crops if there is a substantive relationship to end-use properties of cereal grains Functional and nutritional aspects of cereal-based foods and beverages, whether baked, fermented, or extruded Industrial products (e.g. starch derivatives, syrups, protein concentrates, and isolates) from cereal grains, and their technology
Editor-in-Chief
F. MacRitchie
Grain Science and Industry Department Kansas State University Manhattan, Kansas, USA
Editors
J. Dexter
Grain Research Laboratory Canadian Grain Commission Winnipeg, Canada
G. B. Fincher
Waite Agricultural Research Institute, University of Adelaide, Australia
R. J. Hamer
Wageningen Centre for Food Sciences Wageningen, The Netherlands
D. Lafiandra
Dipartimento di Agrobiologia e Agrochimica University of Tuscia Viterbo, Italy
J. R. N. Taylor
University of Pretoria, South Africa
Reviews Editor
P. R. Shewry
Rothamsted Research Harpenden, UK
Editorial Board
N.-G. Asp University of Lund, Sweden D. E. Briggs University of Birmingham, UK P. Colonna INRA, Laboratoire de Biochimie et Technologies des Glucides Nantes, France K. Denyer John Innes Centre, Norwich, UK P. J. Frazier Dalgety Food Technology Centre, Cambridge, UK R. A. Graybosch University of Nebraska, Lincoln, Nebraska, USA Zhonghu He International Maize and Wheat Improvement Center, Chinese Academy of Agricultural Sciences, Beijing, China G. E. Inglett USDA, National Center for Agricultural Utilization Research, Peoria, Illinois, USA B. O. Juliano Philippines Rice Research Institute, Laguna, Philippines J. Kokini Cook College, Rutgers University, New Brunswick, New Jersey, USA P. Linko Helsinki University of Technology, Espoo, Finland A. W. MacGregor Livingston, Scotland, UK Y. Popineau INRA, Unite de Recherche sur les Proteines Vegetales et leurs Interactions, Nantes, France K. Poutanen VTT Biotechnology, Espoo, Finland S. O. Serna Saldivar ITESM, Monterrey, Mexico B. A. Stone Department of Biochemistry La Trobe University Bundoora, Victoria, Australia B. Svensson BioCentrum-DTU, The Technical University of Denmark, Kgs. Lyngby, Denmark
Founding Editors
T. Galliard and J. D. Schofield Author enquiries
For enquiries relating to the submission of articles (including electronic submission where available) please visit this journals homepage at http://www.elsevier.com/locate/jcs. You can track accepted articles at http://www.elsevier.com/trackarticle and set up e-mail alerts to inform you of when an articles status has changed. Also accessible from here is information on copyright, frequently asked questions and more. Contact details for questions arising after acceptance of an article, especially those relating to proofs, will be provided by the publisher. For a full and complete Guide for Authors, please go to: http://www.elsevier.com/locate/issn/0733-5210/authorinstructions
Journal of Cereal Science 49 (2009) 13
Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/jcs
Rapid Communication
Degradation of cereal beta-glucan by ascorbic acid induced oxygen radicals

Reetta Kivela*, Fred Gates 1, Tuula Sontag-Strohm
Department of Food Technology, University of Helsinki, Agnes Sjoberg Street 2, P.O. Box 66, FIN-00014 Helsinki, Finland
a r t i c l e i n f o
Article history: Received 24 April 2008 Received in revised form 15 August 2008 Accepted 5 September 2008 Keywords: Beta-glucan Degradation Oxygen radicals Ascorbic acid
a b s t r a c t
Degradation of cereal beta-glucan is usually attributed to enzymes or acid hydrolysis. However, there is evidence that polysaccharides are also susceptible to OH-radical induced depolymerisation, and that these radicals can be produced in cereal food systems. The role of Fenton type oxidation was demonstrated in pure beta-glucan solution (0.6%). An addition of ascorbic acid (10 mM) or its oxidation product, dehydroascorbic acid, in the presence of iron sulphate resulted in a signicant decrease of the solution viscosity and molecular degradation of beta-glucan. The viscosity decrease was inhibited by introducing a OH-scavenger (glucose) in the solution or limiting oxygen level in the sample solution. This demonstrates the role of OH-radicals in beta-glucan scission and suggests oxidative cleavage to be a potential threat for the stability of beta-glucan in certain bre enriched products. 2008 Elsevier Ltd. All rights reserved.
We report degradation of cereal beta-glucan in aqueous solutions as a result of a treatment with ascorbic acid and its oxidation product dehydroascorbic acid. It was found that (1 / 3), (1 / 4)beta-D-glucan was depolymerised when treated with ascorbic acid or dehydroascorbic acid and FeSO4 in the absence of enzymatic activity or conditions that could cause acid hydrolysis. The degradation was inhibited by both glucose addition and by displacing oxygen with nitrogen. Non-enzymatic, ascorbate induced degradation of cell-wall polysaccharides is established in vitro and suggested to occur in plant cells during its elongation phase and ripening (Fry, 1998; Fry et al., 2002; Schopfer, 2001). However, this mechanism appears to be neglected in food technology in spite of its potentially crucial role in reducing the physiological efcacy, and hence the validity of health claims, of high beta-glucan foods as well as inuencing properties such as rheology and physical stability. Macromolecules are typically susceptible to degradation induced by reactive oxygen species (ROS). Particularly OH-radicals, one specic species of ROS, have been established to scission polysaccharides (Fry, 1998). One way to produce OH-radicals nonenzymatically is a Fenton type reaction (Reaction (1)). The key feature of Fenton chemistry is the catalytic nature of reduced
Abbreviations: AH2, ascorbic acid; A, dehydroascorbic acid; BBG, beta-glucan isolate from barley; HPSEC, high performance size exclusion chromatography; Mw, weight average molecular mass; OBG, beta-glucan isolate from oat; ROS, reactive oxygen species, i.e. all kinds of chemically reactive molecules derived from ground state dioxygen (3O2). * Corresponding author. Tel.: 358 9 19158540; fax: 358 9 19158460. E-mail address: reetta.kivela@helsinki. (R. Kivela). 1 Present address: Department of Physics, University of Helsinki, P.O. Box 64, FIN-00014 Helsinki, Finland. 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.09.003
metals (Fe2, Cu, Zn), which can produce aggressive OH-radicals from the modestly reactive H2O2 (Wardman and Candeias, 1996). A reducing agent is needed to reduce the catalysing metal to the active state and the dissolved oxygen to H2O2, thus it initiates the Fenton type reaction. All these reagents (dissolved oxygen, transition metal and compounds with reduction potential) are typically present in the aqueous phases of cereal foods, which also contains the soluble bres (Slavin et al., 1999). In addition, beta-glucan is known to form complexes with metals (Platt and Clydesdale, 1984), which makes it a specic target for OH-radical induced degradation, as the short living radicals are produced in immediate vicinity of beta-glucan molecules (Chevion, 1988).
Cu =Fe2 H2 O2 /$ OH OH Cu2 =Fe3
(1)
Ascorbic acid (AH2) is an excellent reducing agent. It readily undergoes a two-step oxidation process forming dehydroascorbic acid (A) and ascorbate radical (AH) as an intermediate. The ascorbate radical is a relatively stable free radical that has a capability of inhibiting oxidation chain reactions and thus can act as a secondary antioxidant. However, ascorbic acid is known to have a dual nature: in certain conditions it behaves as a pro-oxidant instead of antioxidant initiating the Fenton reaction due to the reducing capacity (Reactions (2a) and (2b)). The dual effect is considered to be concentration dependent, so that the rate of antioxidant chain reactions is lowered and the pro-oxidant nature dominates at lower concentrations (Buettner and Jurkiewicz, 1996; Wardman and Candeias, 1996). Ascorbic acid is a common food constituent and commonly added as an antioxidant; it is also used as a our improver which makes Fenton chemistry even more interesting in cereal foods.
R. Kivela et al. / Journal of Cereal Science 49 (2009) 13
AH2 2Cu2 =Fe3 /A 2H 2Cu =Fe2 AH2 O2 /A H2 O2
(2a)
(2b)
Degradation of cereal beta-glucan by the chemical radical reactions may explain some of the controversial results obtained in various studies. It is not uncommon that a reduction in viscosity of beta-glucan (i.e. molecular size reduction) or molecular degradation is observed in food systems such as bakery products and beverages (man et al., 2004; Beer et al., 1997). These changes have generally been attributed to enzymatic or acid hydrolysis or to high shear forces. However, in some cases enzymes seem to have been inactivated in heating processes or the conditions are not suitable for acid hydrolysis to occur, which requires a very low pH (12) and simultaneous heating (Johansson et al., 2006). In addition, there is no clear evidence to link the acid hydrolysis of beta-glucan to the organic acids present in foods. Technologically the degradation of beta-glucan is serious. In aqueous systems the breakdown of the polymer easily leads to product structural breakdown and instability. The degradation of beta-glucan is also nutritionally crucial, since the health benets of beta-glucan are closely related to its rheology. In particular, it is well established that the ability of beta-glucan to attenuate postprandial blood glucose and insulin levels is related to the viscosity and molecular weight of beta-glucan (Wood et al., 2000). The structure of food matrices, in addition to the rheology, is proposed to affect cholesterol lowering of beta-glucan and liquid beta-glucan rich foods have been established to be especially effective (Kerckhoffs et al., 2003; Naumann et al., 2006). There is thus an increasing need to properly understand the threats for instability of native beta-glucan in aqueous environments. In the present study effects of ascorbic acid (10 mM) and dehydroascorbic acid (10 mM) were examined in pure 0.6% betaglucan solutions when added ionic iron (0.01 mM FeSO4$H2O) was present. Also the roles of glucose (1 M, 18 w/w-%) and oxygen were studied. The beta-glucan solutions were prepared from high viscosity barley beta-glucan (BBG, Megazyme, purity >97%) or high viscosity oat beta-glucan (OBG, Megazyme, purity >97%) by wetting dry beta-glucan with ethanol and dispersing the blend in water. The dispersion was rst boiled under stirring for 10 min and then continued for 3 h at 80 C before adjusting the volume in a volumetric ask. Viscosities of the stock solutions were controlled up to three weeks and were observed to maintain constant viscosity. Sodium azide (0.02 w/w-%) was added to prevent microbial enzyme activity. At this concentration, sodium azide was not shown to affect the studied reactions. All the reagents were added as solid compounds to the stock solution. The pH was adjusted to 4.8 0.1 to represent acidic conditions of fermented and beverage-like foods and to enable ascorbic acid to exist in dissociated form (pKaAH2 4.2) and the sample volumes were adjusted after pH adjustment. Viscosity measurements and oscillation tests were performed using a rheometer (ThermoHaake RheoStress 600, Thermo Electron GmbH, Dreieich, Germany). Flow curves were obtained over a shear rate range of 0.3 3000.3 s1 using a double gap geometry (DG41). The mean apparent viscosities of three independent measurements (mean SD) at a shear rate of 10 s1 as a function of time are shown. Dynamic rheological behaviour was characterised by a cone and plate geometry (35 mm/2 ) using a frequency sweep of 0.0110 Hz and a stress of 0.02 Pa. All the rheological experiments were performed at 10 0.1 C. Prior to measurement the samples were stored at 6 0.1 C. The molecular weight of betaglucan was analysed by HPSEC with calcouor-uoresence and dual angle light scattering detection in 0.1 N sodium hydroxide solution (Suortti, 1993).
The rheological studies showed that 10 mM AH2 added together with 0.01 mM FeSO4$H2O caused an approximately 50% drop in viscosity of the barley beta-glucan (BBG) solution in 1 h, while the viscosity of BBG treated with FeSO4 alone remained unchanged (Fig. 1). The results of BBG solution alone are not shown, because the ionic iron treatment did not affect it. Consistent results were obtained with the oat beta-glucan solution (OBG, results not shown). The viscosity behaviour of these solutions changed from shear thinning to Newtonian in 2 h suggesting that the critical concentration, at which the entanglements between molecules or aggregates are lost, had been reached as a result of molecular size reduction (Fig. 2). HPSEC-studies showed that this reduction in molecule size was depolymerisation instead of other changes in physicochemical properties, since the weight average molecular mass (Mw) of beta-glucan was reduced from 520 (5) to 35 (2) 103 g/mol as a result of the ascorbic acid and iron treatment. The effect of the third reagent, oxygen, was demonstrated by limiting the amount of dissolved oxygen with nitrogen supplement during sample preparation and testing. After 4 h of N2-treatment, air was mixed into the nitrogen treated ascorbic acid-BBG sample in the measuring cap of the rheometer and this caused a drop in viscosity (Fig. 1). Thus all the reagents of Fenton chemistry, i.e. transition metal, dissolved oxygen and a reducing agent (Reactions (1)(2b)), were shown to be needed for the depolymerisation of beta-glucan in aqueous solution. Dehydroascorbic acid (10 mM) caused a 30% viscosity reduction when added with ionic iron to the beta-glucan solution (Fig. 1). The total decrease during the measuring time was appr. 50%, while ascorbic acid decreased the viscosity over 90%. The lower viscosity reduction, i.e. the rate of oxidative cleavage compared to the action of ascorbic acid agrees with the results of Fry (1998) with xyloglucan. Dehydroascorbic acid is likely to occur in food systems, since ascorbic acid undergoes a reversible oxidation reaction induced by light, heat and metal rst forming semidehydroascorbic
180
Apparent viscosity at10 1/s (mPas)
160 140 120 100 80 60 40 20 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time / h
+1M glucose Control +10 mM AH2+ 1M glucose +10 mM A +10 mM AH2, limited oxygen +10 mM AH2
Fig. 1. The effect of ascorbic acid (AH2) and dehydroascorbic acid (A) with ionic iron on the viscosity of beta-glucan (control, BBG, 0.6%) and the inhibition effects of oxygen limitation and glucose addition (1 M). The arrow shows the point of air introduction into the nitrogen treated sample. All the samples contained 0.01 mM FeSO4$7H2O.
R. Kivela et al. / Journal of Cereal Science 49 (2009) 13
0,1
Control 0.5 h 1h 2h
0,01
4h 6h 8h 10 h 18 h 1 10 100 1000
the reagents, i.e. a reducing agent (ascorbic acid), transition metal (iron) and oxygen, were shown to have a role in the depolymerisation of beta-glucan. The viscosity loss was inhibited by an oxygen radical scavenger (glucose), which supports the role of OH-radicals in the non-enzymatic cleavage of cereal beta-glucan. Common food additives as well as the minor compounds of the soluble bre phases may be harmful for the stability of aqueous beta-glucan enriched foods. In addition, their role in the research of health benets of beta-glucan in different processed foods needs proper understanding. Acknowledgements The authors thank Dr. Tapani Suortti (VTT, Finland) for performing the molecular weight measurement and data analysis and Professor Hannu Salovaara and Dr. Laura Nystrom for their important support in the process. References
Autio, K., Myllymaki, O., Malkki, Y., 1987. Flow properties of solutions of oat bglucans. Journal of Food Science 52, 13641366. Beer, M.U., Wood, P.J., Weisz, J., Fillion, N., 1997. Effect of cooking and storage on the amount and molecular weight of (1 / 3), (1 / 4)-b-D-glucan extracted from oat products by an in vitro digestion system. Cereal Chemistry 74, 705709. Buettner, G.R., Jurkiewicz, B.A., 1996. Catalytic metals, ascorbate and free radicals: combinations to avoid. Radiation Research 145, 532541. Buxton, G.V., Greenstock, C.L., Helman, W.P., Ross, A.B., 1988. Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (OH/O) in aqueous solution. Journal of Physical and Chemical Reference Data 17, 513769. Chevion, M., 1988. A site-specic mechanism for free radical induced biological damage: the essential role of redox-active transition metals. Free Radical Biology 5, 2737. Fry, S.C., 1998. Oxidative scission of plant cell wall polysaccharides by ascorbateinduced hydroxyl radicals. Biochemical Journal 332, 507515. Fry, S.C., Miller, J.G., Dumville, J.C., 2002. A proposed role for copper ions in cell wall loosening. Plant and Soil 247, 5767. Johansson, L., Virkki, L., Anttila, H., Esselstrom, H., Tuomainen, P., Sontag-Strohm, T., 2006. Hydrolysis of b-glucan. Food Chemistry 97, 7179. Kerckhoffs, D., Hornstra, G., Mensink, R.P., 2003. Cholesterol-lowering effect of betaglucan from oat bran in mildly hypercholesterolemic subjects may decrease when beta-glucan is incorporated into bread and cookies. The American Journal of Clinical Nutrition 78, 221227. Lazaridou, A., Biliaderis, C.G., Izydorczyk, M.S., 2003. Molecular size effects on rheological properties of oat beta-glucans in solution and gels. Food Hydrocolloids 17, 693712. Morelli, R., Russo-Volpe, S., Bruno, N., Lo Scalzo, R., 2003. Fenton-dependent damage to carbohydrates: free radical scavenging activity of some simple sugars. Journal of Agricultural and Food Chemistry 51, 74187425. Naumann, E., van Rees, A.B., Onning, G., Oste, R., Wydra, M., Mensink, R.P., 2006. Beta-glucan incorporated into a fruit drink effectively lowers serum LDLcholesterol concentrations. The American Journal of Clinical Nutrition 83, 601 605. Platt, S.R., Clydesdale, F.M., 1984. Binding of iron by cellulose, lignin, sodium phytate and beta-glucan, alone and in combination, under simulated gastrointestinal pH conditions. Journal of Food Science 49, 531535. Schopfer, P., 2001. Hydroxyl radical-induced cell-wall loosening in vitro and in vivo: implications for the control of elongation growth. The Plant Journal 28, 679 688. Slavin, J.L., Martini, M.C., Jacobs, D.R., Marquart, L., 1999. Plausible mechanisms for the protectiveness of whole grains. The American Journal of Clinical Nutrition 70, 459463. Suortti, T., 1993. Size-exclusion chromatographic determination of beta-glucan with postcolumn reaction detection. Journal of Chromatography 19, 105110. Tosh, S.M., Wood, P.J., Wang, Q., Weisz, J., 2004. Structural characteristics and rheological properties of partially hydrolyzed oat b-glucan: the effects of molecular weight and hydrolysis method. Carbohydrate Polymers 55, 425436. Wardman, P., Candeias, L.P., 1996. Fenton chemistry: an introduction. Radiation Research 145, 523531. Wood, P.J., Beer, M.U., Butler, G., 2000. Evaluation of role of concentration and molecular weight of oat beta-glucan in determining effect of viscosity on plasma glucose and insulin following an oral glucose load. The British Journal of Nutrition 84, 1923. man, P., Rimsten, L., Andersson, R., 2004. Molecular weight distribution of bglucan in oat-based foods. Cereal Chemistry 81, 356360.
Viscosity Pas
0,001 0,1
Shear rate 1/s

Fig. 2. Flow curves of ascorbic acid (10 mM) treated beta-glucan solution (0.6%, BBG) from 0 h to 18 h and the control at 18 h. Both samples contained 0.01 mM FeSO4$7H2O.
(AH) acid and further dehydroascorbic acid (A). It has been reported that, like ascorbic acid, dehydroascorbic acid can also keep metals such as copper in their reduced form, but it cannot reduce oxygen to H2O2 (Fry, 1998). Yet as endogenic H2O2 usually occurs in water solutions and especially in natural solutions such as food systems, dehydroascorbic acid may also participate in the oxidative cleavage of beta-glucan. Glucose efciently inhibited the ascorbic acid induced viscosity reduction of BBG solution (Fig. 1). Glucose was introduced to the study as a competitive OH-radical scavenger (Buxton et al., 1988). In the solutions with high concentration of glucose (1 M, appr. 20 w/ w-%), the majority of indiscriminant OH-radicals will react with the glucose rather than with the high molecular weight polysaccharide (beta-glucan, 0.6 w/w-%). After the reaction with OH, glucose itself becomes a relatively stable free radical. In addition, the glucose radical may regenerate itself with a hydrogen atom from the solution, thus enhancing its stability and efciency as a scavenger (Morelli et al., 2003). The viscosity of glucoseBBG solution was 17% higher than the viscosity of BBG solution alone, which is in accordance with that reported before (Autio et al., 1987), and it decreased 20% due to the ascorbic acid treatment (Fig. 1). However, the viscosity of ascorbic acidglucoseBBG solution remained unchanged as a function of time indicating that the proposed OHinduced degradation of beta-glucan was inhibited by glucose addition. The ow curves of ascorbic acid treated BBG solution showed hysteresis and noise after 5 h (Fig. 2). This may both indicate novel networks and particle formation. Gel formation was observed as increased turbidity in the relatively dilute ascorbic acidBBG solution (0.6%) after four days storage, while the control BBG solution stayed clear. The gelation was shown by dynamic rheology, the elastic modulus (G0 ) being nearly frequency-independent (ca. 10 Pa at 1 Hz) and the loss modulus (G00 ) being low (appr. 1 Pa at 1 Hz) at frequencies 110 Hz. It is known that gelation rate as well as gel strength increase with decreasing molecular weight so that molecules with an Mw above 250 000 g/mol do not practically form gels (Lazaridou et al., 2003; Tosh et al., 2004). Such oxidative cleavage may bring new aspects for the phenomenon of gelation during food processing, since it can occur in enzyme and acid hydrolysis free conditions and the cleavage may result in different compounds as found in hydrolysis (Fry, 1998; Fry et al., 2002). In conclusion, it was found that ROS represents a threat to the stability of cereal beta-glucan in an aqueous environment, and that Fenton chemistry results in depolymerisation of beta-glucan. All

Variation in Granule Bound Starch Synthase I (GBSSI) loci amongst Australian wild cereal relatives (Poaceae)
F.M. Shapter*, P. Eggler, L.S. Lee 1, R.J. Henry
Grain Foods CRC, Centre for Plant Conservation Genetics, Southern Cross University, PO Box 157, Lismore, NSW 2480, Australia
Article history: Received 30 July 2007 Received in revised form 12 November 2007 Accepted 20 June 2008 Keywords: Australian native grasses Starch Crop wild relatives Cereal Phylogenetic
a b s t r a c t
A complex cascade of enzymes is responsible for the development of starch granules in grain endosperm. Granule Bound Starch Synthase I (GBSSI), encoded by the Waxy gene, is a key enzyme of starch synthesis and determines the accumulation of amylose in the starch granules. The complete genomic GBSSI sequence was ascertained for eight Australian cereal wild relatives (CWR) to determine diversity within the gene. A phylogeny derived from the coding sequence of the entire Waxy gene was compared to established phylogenetic relationships. Starch granule morphology observed in conjunction with this phylogeny suggests that small polygonal starch granules arranged as compound granules are the ancestral state, evolving subsequently to bimodal starch granules and to larger simple granules. Genomic sequence length varied within the species from 2800 to 3572 bp. Most variation occurred within the intron sequences, the largest insertion showing strong homology to a retrotransposon. One wild species was determined to have a deletion in the 30 -end of exon 1 resulting in a putatively non-functional allele. Alignment of the amino acid sequence showed strong homology throughout the central fragments of the gene but broad variation in the transit peptides. All putative functional alleles maintained the reported active sites for glycogen synthesis, though with variations in other highly conserved areas of the gene. These variations within the wild relatives of cultivated cereals may provide novel sources of genetic diversity for future cereal improvement programs. 2008 Elsevier Ltd. All rights reserved.
1. Introduction Granule Bound Starch Synthase I (GBSSI), encoded by the Waxy gene, is responsible for the accumulation of amylose during the development of starch granules in cereal endosperm. GBSSI has been reported with changing enzyme nomenclature over the last two decades. The most recent, EC 2.4.1.242, clearly delineates this enzyme from previously recorded enzymes EC 2.4.1.11 (Taira et al., 1995) and EC 2.4.1.21 (Baldwin, 2001; Demeke et al., 1999), as being able to use both UDP-glucose and ADP-glucose as a substrate for starch synthesis. Although GBSSI and II share the same enzyme nomenclature and approximately 69% sequence identity, suggesting that they are homologous, the genes encoding them are situated at independent loci. The gene encoding GBSSI is predominantly expressed in endosperm and pollen whereas GBSSII is expressed in the leaf, culm and pericarp (Vrinten and Nakamura, 2000). Equally well reported is that cereal species with a null allele at all Waxy loci will produce an amylose free endosperm. The
* Corresponding author. Tel.: 61 2 6620 3466; fax: 61 2 6622 2080. E-mail address: frances.shapter@scu.edu.au (F.M. Shapter). 1 Present address: Centre for Tropical Crops and Biocommodities, Queensland University of Technology, GPO Box 2434, Brisbane, Queensland 4001, Australia. 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.06.013
amylose status of cereal starches affects their functional properties and therefore their end uses (Domon et al., 2002; Gaines et al., 2000; Kim et al., 2003) and a simple, rapid iodine test for the presence/absence of amylose in endosperm has been recently optimised (Pedersen et al., 2004). Within cultivated cereals Waxy mutants are common from induced (Nakamura et al., 1995) and naturally occurring origins (Domon et al., 2002), with mutations remaining conserved in the genome over generations. The concentration of Waxy mutants occurring in cereal wild relatives (CWR) is as yet undetermined. Of the waxy and low-amylose cereal species sequenced to date, it appears that there is no specic polymorphism across all the species which results in the waxy phenotype. Rather, any disruption to the production or regulation of this enzyme is critical in cereal species (Mangalika et al., 2003; Rahman et al., 2000; Shure et al., 1983; Yan et al., 2000). Within any one cereal species, multiple alleles of the GBSSI coding gene are common and are denoted by a subscript or chromosomal reference which is species specic (Mangalika et al., 2003; Pedersen et al., 2007). In the case of polyploid species such as wheat, the number of functional GBSSI coding alleles is correlated with the concentration of amylose in the grain (Mangalika et al., 2003). The phylogenetic utility of the GBSSI encoding gene has been repeatedly reported (Evans and Campbell, 2002; Mason-Gamer
F.M. Shapter et al. / Journal of Cereal Science 49 (2009) 411
et al., 1998; Small et al., 2004). Phylogenetic relationships between species provide an understanding of the breadth of their genetic diversity and have been utilised in crop-breeding programs to determine relationships between cultivated species and their wild cereal relatives (Williams et al., 2006). Initial phylogenetic analyses across this subset of Poaceae utilising a small central fragment of the gene was insufcient to resolve the basal bifurcations and interspecies relationships (data not shown). As a result, more sequence was captured to enhance phylogenetic analysis. In order to ensure that phylogenetic predictions are resolving species relationships and not gene evolution, analysing more than one gene concurrently is prudent (Small et al., 2004). However, functional genes evolve under selective pressure and therefore their evolutionary histories can be correlated with their functional evolution (Ochieng et al., 2007). In this instance establishing the diversity and evolution of GBSSI is the primary objective. Following crop improvements made by plant breeders in the last century, cereal-breeding programs are increasingly targeting cereal wild relatives as a source of novel germplasm (Hajjar and Hodgkin, 2007). This is occurring in a time of increasing human populations and climatic instability. It has been proposed that meeting global food requirements by 2025 will require a 25% increase in grain production (Khush, 2001). The use of CWR in plant breeding programs will become increasingly important because of their environmental adaptations (Ashikari and Matsuoka, 2006). Previously, hybridisation of CWR from the primary gene pool of the respective cultivated cereals has been used to improve agronomic traits such as growth habit, and pest and disease resistance. Recent advances in molecular and plant breeding techniques have shifted the focus from primary to secondary and/or tertiary gene pool species as sources of novel allelic variation for genes of importance (Hajjar and Hodgkin, 2007; Rao et al., 2003). Characterisation of the full coding region of GBSSI in selected native species may identify variants of the GBSSI coding gene. In this study GBSSI gene sequence is reported, and used as an indicator of the breadth of genetic diversity within and between Australian CWR and cereals. Furthermore, alignment of the amino acid sequences derived from the genomic data generated identies novel putative GBSSI amino acid sequences in a subset of the Australian CWR. 2. Materials and methods 2.1. Plant material and DNA extraction Representative species of the four major clades of Poaceae found in Australia were selected for their close relationship to a cereal species, large seed size, broad geographic distribution and/or, in the case of Microlaena stipoides and Astrebla lappacea, their proven potential to be cultivated. Seed was germinated and leaf material harvested for DNA extractions. Species identity was conrmed by a herbarium and voucher details are listed in Table 1. Field collection was undertaken in Northern NSW for A. lappacea and M.
stipoides. Elymus scaber and Austrostipa aristiglumis were sourced from Native Seeds Pty Ltd Australia (http://www.nativeseeds.com. au). Native Oryza and Sorghum seeds were sourced from the Australian Tropical Crops and Forages Collection, Queensland Department of Primary Industries and Fisheries, Biloela, Australia (www.dpi.gov.au/auspgris/). DNA was extracted using a Qiagen DNeasy Plant Maxi Kit to ensure sufcient high-quality/highconcentration DNA from a single extraction. DNA quality was conrmed by gel electrophoresis and quantied using a NanoDrop Technologies ND-1000 spectrophotometer. 2.2. Presence of amylose The presence of amylose was determined by Pedersons iodine staining technique (Pedersen et al., 2004). Controls for waxy (Sorghum bicolor cv. Batsuto Red) and non-waxy (S. bicolor cv. LR2409) endosperm (Pedersen et al., 2007) were used. 2.3. Sequence alignment GBSSI coding DNA sequence of Oryza sativa (X65183), Triticum aestivum (AB019623), Hordeum vulgare (AB089162), Zea mays (X03935), Setaria italica (AB089141) and S. bicolor (AF488412.1) was retrieved from the NCBI Genbank Database (http://www.ncbi.nlm. nih.gov/). Sequence alignments were performed using ClustalW (EMBL European Bioinformatics Institute; http://www.ebi.ac.uk/ clustalw/). 2.4. PCR amplication and DNA sequencing Universal primer pair C and E from McIntosh et al. (2005) was used in accordance with the described method, to amplify a small central fragment (CE) of the GBSSI gene in the wild grasses. Small fragment PCR products were puried using a Qiagen Gel Extraction Kit prior to DNA sequencing. Full gene amplication was obtained by PCR using Accuprime Hi-Fi (Invitrogen) according to the manufacturers guidelines. To increase specicity, DMSO and 50% glycerol were added to the PCR reaction at 3 and 6%, respectively, for some of the species (Table 2). Heteroduplexes can be induced by PCR of multi-copy loci, followed by mismatch repair during cloning (Jansen and Ledley, 1990; Jensen and Straus, 1993; Longeri et al., 2002). To ensure sequence integrity, reconditioning PCRs were performed prior to cloning of the full gene fragments (Thompson et al., 2002). Reconditioning PCRs were spiked with 10% volume of the primary PCR product as template; otherwise PCR conditions were identical to that of the original PCR. Sequencing reactions utilised ABI PRISM Big Dye Terminator (BDT) V3.1 chemistry at one-eighth concentration reactions and the resulting amplicons analysed by capillary electrophoresis on an ABI 3730 Genetic Analyzer by Southern Cross Plant Genomics (SCPG, Lismore, Australia). Sequence data were
Table 1 Classication, collection record, herbarium voucher and Genbank accession number of eight native grass species included in this study Species Astrebla lappacea Austrostipa aristiglumis Elymus scaber Microlaena stipoides Oryza australiensis Oryza rupogon Sorghum leiocladum Sorghum nitidum Tribe Cynodonteae Stipeae Triticeae Erharteae Oryzeae Oryzeae Andropogoneae Andropogoneae Collection recorda AC04-1003495 AC04-1003487 AC04-1003502 AC04-1003504 AusTRCF310676 AusTRCF 309313 AusTRCF 300187 AusTRCF 302543 Herbarium voucher BRIAQ 751244 BRIAQ 751222 BRIAQ 751227 BRIAQ 751226 BRIAQ723945 BRIAQ723944 DNAD0155695 CANB479881 Genbank accession EF600041 EF600042 EF600043 EF600044 EF600036 EF600037 EF600038 50 EF600039; 30 EF600040
a Collection numbers prexed with AC04 can be located through the Australian Plant DNA Bank (www.dnabank.com.au) and those prexed with AusTRCF through the Australian Plant Genetic Resources Information Service (www.dpi.qld.gov.au/auspgris/).
Table 2 Full length gene primers and PCR conditions for A. aristiglumis (Aa), A. lappacea (Al), E. scaber (Es), S. leiocladum (Le), M. stipoides (Ms), O. australiensis (Oa) and O. rupogon (Or) Primer name GBSSAaF1 GBSSAaR1 GBSSAlF3 GBSSAlR2 GBSSEsF3 GBSSEsR3 GBSSLeF5 GBSSLeR5 GBSSMsF1 GBSSMsR2 GBSSOaF1 GBSSOaR2 GBSSOrF1 GBSSOrR2 Sequence 50 30 AGTGCAGTCATCTTTCACCA CAGAACCTCAAACTTATTAGCC ACCGCAGCTACGTACTCCGCC CACCGATCCCCCTTGTTCT TGAGGATTCATGCTCTAGGGTTA AAGAGGCAAGCGGCACA GCACATGTCTTTTCTTGATGC TGATCCATCATCCTTGTGCTA AAAATTCTGTTATCCCAACCC ATAATGGCTACGACAGACTCAC ACAGCAAGAGCTAGACAACCG AAACCACTGCTTCATTCGTCT ATCTTCCACAGCAACAGCTA ATCACTCCATATAGATCTCAGGC Tm ( C) 53.1 53.6 60.0 59.3 58.6 57.8 55.7 56.1 55.4 53.9 57.4 56.6 53.1 54.8 Size (bp/mer) 20 22 21 19 23 17 21 21 21 22 21 21 20 23 GC (%) 45.0 40.9 67.7 57.9 43.5 58.8 42.9 42.9 39.1 45.5 52.4 42.9 45.0 43.5 PCR Step 1a (cycle #/temp. C) 8/5850 10/6253 12/6048 12/6048 8/850 8/5850 10/6050 PCR Step 2a (cycle #/temp. C) 30/50 30/52 30/49 30/49 30/50 30/50 30/50 Polymerase Accuprime DG Accuprime DG Accuprime Accuprime Accuprime DG Accuprime DG Accuprime
a Two step PCR was utilised: the rst step a touchdown cycle, each cycle the annealing temperature decreasing 1 C within the given range, the second step at a constant annealing temperature for the given number of cycles.
aligned and edited using Sequencher V4.5 (Gene Codes Corporation, Ann Arbor, USA). 2.5. Genome walking Species-specic primers were designed using Primer Premier V5.0 (Premier Biosoft International, Palo Alto, USA). To amplify the gene from the central CE fragment out to the 50 - and 30 -UTR, respectively, the BD GenomeWalker Universal Kit (BD Biosciences, San Jose, USA) was used, as per the manufacturers instructions, except that all PCR amplication mixes were prepared at half volume. A mixture of 3% DMSO and 6% glycerol was added to the PCR mix and ve cycles added to the standard PCR program, as per the manufacturers trouble-shooting guide. 2.6. Cloning and DNA sequencing GenomeWalker and full gene PCR products were run on a 1% low-melt agarose gel and the individual PCR product bands of interest excised. Immediately prior to ligation into the vector, the gel slice was melted in a thermocycler at 72 C for 1 min and then held at 37 C. Three microliter of molten gel was included in the standard ligation reaction mix of the pGEM-T Easy Vector II System (Promega Corporation, Madison, USA) following the manufacturers recommended protocol. Clones were screened for size by PCR using Roche Taq (Roche Products Pty Limited, Australia) in the manufacturers recommended reaction mix. Target clones were isolated, then amplied using the Templiphi System (GE Healthcare) as per the manufacturers instructions. Standard BDT V3.1 one-eighth sequencing reactions used 5 ml of diluted Templiphi as template. 2.7. Exon/intron boundaries The putative location of each exon/intron boundary was determined by alignment with cDNA sequence from Genbank accessions of the CWRs most closely related available cDNA. The start of the introns was then assigned as being the nearest GT to that position and ending on the most distal AG prior to the following exon, whilst maintaining the open reading frame (Murai et al., 1999; Ramakrishna et al., 2002). 2.8. Species comparisons Neighbour-joining phylogenetic analyses were conducted in PAUP*Version 4.0b10 (Swofford, 2001), using Kimura-2 parameter,
distance method and maximum parsimony settings. Support for branches was obtained from 1000 pseudoreplicates for 1920 of 1921 characters, using the TBR branch swapping algorithm.
3. Results 3.1. Gene sequence Initial alignment of the GBSSI coding sequence from Genbank determined that sites for universal primer design at the 50 - and 30 ends of the gene were not available due to poor homology between the species in these regions. Direct sequencing analysis of puried PCR amplicons of a small central fragment (CE) of the GBSSI gene spanning exons 7 and 8 (McIntosh et al., 2005) produced poor sequence data for some species with ambiguous base calls indicating multiple alleles. This was expected as, apart from the two wild Oryza species, the CWR are polyploids and hence sequence data could only be resolved by cloning the PCR products prior to sequencing. Fragment CE sequence was used as a platform for the design of bidirectional species-specic primers for use in the BD Genomewalking protocol. The rst set of gene/species-specic primers produced gene fragments that did not reach the start and stop codons. Up to three sets of gene/species-specic primers were designed in the exon closest to the respective 50 - and 30 -ends of the sequences obtained in the current round of genome walking (Fig. 1). Addition of the recommended concentrations of DMSO and glycerol in the second and third rounds of genome walking resulted in improved amplication specicity. Of the eight species selected for full gene characterisation, seven were successfully amplied with full gene primers located in the 50 and 30 -UTR. Due to the polyploid status of most of these species it was necessary to amplify the entire gene sequence in a single PCR fragment, then clone the amplicon. This ensured the sequence was obtained from a single genome. In the eighth species, Sorghum nitidum, genome walking produced two major gene fragments, overlapping at intron 4 and the start of exon 5. Polymorphisms between the overlapping fragments determined that these were two different alleles; indicating more than one genome may have been amplied. Although more than 20 primer pairs were utilised under various PCR conditions, full gene amplication was unsuccessful for this species. Consequently, S. nitidum has been entered into Genbank as two separate (50 and 30 ) sequences (Table 1). To determine a putative coding sequence for S. nitidum, the gene sequence was spliced, in silico, at
Fig. 1. Schematic diagram showing/illustrating the variations observed across eight Australian grass species and their cultivated cereal relatives. The location of central fragment (CE) and the genome-walking gene/species-specic primers are indicated for each species.
exon 5 and degenerate base calls inserted for the three polymorphic positions in the overlapping fragment. A. lappacea appeared to be missing the 30 portion of exon 1 and missing its expected exon 1/intron 1 boundary when compared to all other species examined. Exons 2 and 3 were as expected; however, the 30 boundary of exon 4 occurred 1 base earlier, causing a frame shift when translated. Other than exons 5 and 11, the exon sizes varied by between 1 and 29 bases, causing repeated frame shifts and encoding multiple stop codons commencing with the rst in exon 5. Amongst all the species examined, the genomic base count ranged from 2760 bases (E. scaber) through to 3572 bases (Oryza rupogon). This large variation in nucleotide number can be accounted for in the Triticeae by the absence of introns 4 and 7. The insertion of over 400 bases in intron 11 of O. rupogon is partially responsible for its larger size (Fig. 1). The number of putative coding bases varies from 1812 base pairs (bp) in A. aristiglumis up to 1830 bp for both Oryza australiensis and O. rupogon, with the four cereal species investigated falling within this range. This translates to an addition/deletion of up to 14 amino acids. 3.2. Transposable elements
3.3. Granule Bound Starch Synthase I Comparison of the putative amino acid sequences derived in this study with those for the cereals retrieved from Genbank showed little conservation within the 79 amino acid consensus positions of the transit peptide (Fig. 2). The consensus of the mature protein was 540 amino acids and averaged 75% identity. The conserved domain database (Marchler-Bauer et al., 2007) indicates two conserved domains exist in GBSSI; the glycosyl transferase groups ve and one. Group ve has 79.5% identity and group one maintains 71.4% identity. The C-terminus has 74.7% identity within the amino acid sequence. The ADPG active binding site near the N-terminus (Baldwin, 2001; Denyer et al., 2001) was identical in all species except A. lappacea and S. nitidum. A single amino acid change occurred in the spliced S. nitidum sequence. The deletion of part of exon 1 in the A. lappacea sequence translated to the loss of the entire binding site. Of the two conserved motifs near the C-terminus (Baldwin, 2001), Motif II had a proline substituted for the threonine in the two Oryza species. Motif III was conserved across all species.
3.4. Phylogenetic outcomes In Sorghum leiocladum the rst set of genome walking gene/ species-specic primers produced a 1300 bp fragment. Nine hundred bases of this fragment were homologous to exons 3 to 7 of the O. sativa GBSSI coding sequence. At the 50 -end of the fragment the rst 400 bases after the GenomeWalker AP2 primer sequence had no homology to the exons of any GBSSI coding gene in Genbank. These 400 bases were isolated and submitted to a BlastN search which returned a strong identity to a retrotransposon isolated from S. bicolor. Similarly, PCR amplication within the O. australiensis GenomeWalker library produced a fragment which had the GenomeWalker AP2 primer sequence at both ends of the fragment. The sequence showed no homology to the cultivated rice GBSSI exons, and was determined by BlastN search to be homologous to an O. australiensis retrotransposon. The 400 base insert in intron 12 of O. rupogon did not return any matches when used as a BlastN query sequence. Phylograms were developed containing all the possible boundary variations for the A. lappacea sequence (data not shown), including a sequence with no intron 1. All 16 permutations were monophyletic and remained a sister group to the remainder of the Poaceae examined. The representative coding sequence of A. lappacea used in the phylogenetic analyses was determined by having the least variation in exon sizes and the least amino acid sequence disruptions. Preliminary inclusion of the spliced S. nitidum sequence resulted in its position within the topology changing when analyses between the second half of the gene sequence were compared to the full coding sequence. Using the entire sequence S. nitidum was placed between Z. mays and S. italica. When only the second half of the gene was analysed S. nitidum was clustered with S. leiocladum with stronger statistical support. As S. nitidum sequence data was not derived from
Fig. 2. Amino acid alignment of GBSSI and associated conserved domains.
a single genome it was excluded from the phylogenetic analysis presented. No suitable outgroup was available (from Genbank) for the phylogenetic analysis so all trees were unrooted. Neighbourjoining and maximum parsimony analyses of the full coding sequence produced trees with A. lappacea as an outgroup to the
rest of the Poaceae examined (Fig. 3). Analyses produced a single congruent phylogeny with good bootstrap support except at the bifurcation of A. aristiglumis and M. stipoides which had lower bootstrap values. The basal bifurcation of the trees separates the BEP (Bambusoideae, Erhartoideae and Pooideae) from the PACCAD (Panicoideae, Arundinoideae, Chloridoideae,

Alappacea
Oaustraliensis 100 100 100 61 60 100 Osativa Orufipogon
Compound granules
63 66
Mstipoides
Aaristiglumis 51 53 100 100 71 75 Hvulgare
Escaber
T.aestivum
Bimodal granules
Sleiocladum 92 91 100 100 100 100 Sbicolor
Simple granules
Zmays
Sitalica 50 changes
Compound granules
Fig. 3. Neighbour-joining Tree of GBSSI complete coding sequence. Bootstrap values are calculated with 1000 replicates with neighbour-joining values presented above the branches and maximum parsimony values below. The type of starch granule morphology reported for each species is noted on the right of the phylogram.
Centothecoideae, Aristidoideae and Danthoniodeae) clades. Previous phylogenetic analyses utilising the Waxy gene did not incorporate the entire coding region and therefore a second tree was produced excluding characters 1986 which resulted in no changes to the tree topology. 3.5. Iodine test The iodine test for the presence or absence of amylose determined that all eight of the CWR had amylose present in their endosperm (Fig. 4). 4. Discussion GBSSI was successfully amplied in seven of the eight species examined. For the eighth species, S. nitidum, screening multiple genome walker libraries was unable to isolate the same locus for the 50 - and 30 -gene fragments. As such the two overlapping sections of the gene that were sequenced did not provide matching UTR
Fig. 4. Iodine test for the presence or absence of amylose. A magenta colour indicates a waxy phenotype (BR) and blue indicates the presence of amylose (LR). Aa A. aristiglumis, Al A. lappacea, Es E. scaber, Ms M. stipoides, Le S. leiocladum, Ni S. nitidum, Oa O. australiensis, Or O. rupogon.
sequence for full gene primers to be successfully applied. Comparison of the UTR of the other CWR and cereals sequenced showed no homology in this region even amongst closely related species. It is therefore not surprising that the sequence differentiation between genomes in this region of the gene is diverse enough to prevent allele specic primers from amplifying. The GBSSI homolog amplied in A. lappacea was missing a signicant portion of exon 1 which translated to a putative active site in the protein. There was little conservation of the exon/intron boundaries throughout the gene and once translated, multiple stop codons were identied throughout the transcript. It is proposed that this sequence was derived from a non-functional ortholog. Null-alleles of this gene in diploids result in a waxy or amylose free phenotype. For example, an indigenous barley mutant with a large deletion including exon 1 was determined to have a waxy phenotype (Domon et al., 2002). Testing for amylose indicated that all eight CWR contained some amylose (Fig. 4) in their endosperm. The ploidy of A. lappacea is not clearly stated in the literature, though its base chromosome number of 2n 40 (Jozwik, 1969) would suggest some level of polyploidy. It can be assumed then, that A. lappacea has another GBSSI locus yet to be characterised, that is functional and being expressed in the endosperm. Wheat lines with multiple GBSSI loci still produced amylose in the endosperm in lower concentrations than normal when null-alleles did not occur concurrently at all GBSSI loci (Mangalika et al., 2003). Transposable elements have been previously identied in the intron sequence of the Waxy gene (Mason-Gamer et al., 1998; Wang et al., 1994). In the case of S. italica, insertion of transposable elements in the coding region is responsible for a waxy phenotype (Kawase et al., 2005). Multiple attempts to design genome-walking primers to amplify past the retrotransposon in the 50 -fragment of S. leiocladum were undertaken, though remained unsuccessful. It was hypothesised that the retrotransposon may be too large to amplify across in a single genome walking reaction. An alternative genomewalking library was used to amplify the 50 -end of the gene from another genome in which no retrotransposon was present in intron 3. Retrotransposons have been reported as being specic within a genus or tribe (Bennetzen et al., 2004). Finding retrotransposable elements within the CWR genomes is to be expected as these transposable elements can comprise up to 50% of nuclear DNA content (Kumar and Bennetzen, 1999). The inclusion of transposable elements has also been associated with the larger genome size of O. australiensis (Vaughan et al., 2003). Sequence alignment of the complete gene sequence from multiple species identied hot spots of coding sequence divergence between the species. The rst 200 and nal 100 bp of the GBSSI coding gene were particularly poorly conserved within Poaceae. Amongst the putative functional alleles, variable intron sizes (Fig. 1) indicating insertions/deletions were observed, across the species with especially large variations within intron 12. Phylogenetic investigation used these indels to identify homoeologous copies of the Waxy gene (Fortune et al., 2007). The gene variation observed in the current study concurs with comparative studies of other homologous grass genes, where exon/intron boundaries remain conserved, but the size of the introns varies (Feuillet and Keller, 2002). The coding sequence also showed considerable diversity across the taxa sampled. Early protein studies determined that the molecular weight of GBSSI varied from 58 to 60 kDa between species (Taira et al., 1995). Amongst the CWR, coding sequence diversity translates to the omission of 14 amino acids over the length of the protein and multiple polymorphisms in the amino acid sequence. Any changes to the amino acid sequence have the potential to cause novel phenotypic expression of this gene. Previous studies of diploid mutants containing a null Waxy allele revealed GBSSI to be responsible for amylose synthesis in the storage tissues of plants
10
(Vrinten and Nakamura, 2000). Sequence variations, which translate to amino acid changes, may lead to the synthesis of novel starches in planta, thus creating the potential for plant breeders to manipulate starch synthesis through conventional breeding or transgenic modication of cereal crops. The production of cereals containing novel starches will reduce or eliminate the need for costly post-harvest modications (Davis et al., 2003). Introgression of novel genes from CWR has already been responsible for major cereal production gains (Hajjar and Hodgkin, 2007; Langridge et al., 2001). The allelic variation observed in this small subset of Australias native grasses shows that considerable functional diversity may exist in the native grass gene pool, which could be harvested for grain improvement programs. The basal bifurcation of the tree separates the BEP and PACCAD clades, as was reported from previous phylogenetic studies of the Poaceae (Kellogg, 2002; Preston and Kellogg, 2006; Tomlinson and Denyer, 2003). Mason-Gamer et al. (1998) concluded that sequencing through the introns of GBSSI was unnecessary to resolve relationships between genera, tribes and subfamilies within Poaceae. Sufcient resolution was obtained from cDNA data and the results from the current study support their ndings. Similarly, Mason-Gamer et al. (1998) utilised just the 30 half of the gene, which can be universally amplied for analysis. Comparisons of phylogenies developed from this smaller region of the gene compared to the full coding sequence did not result in any change in topology. This analysis indicates that time consuming and costly acquisition of sequence data from non-conserved areas of the Waxy gene does not provide any clearer resolution of phylogenetic inferences for the Poaceae. A previous scanning electron microscopy study (Shapter et al., 2008) identied the starch granule morphology of the eight species and their cereal relatives. Since these trees were developed using the entire coding region of the Waxy gene, and given the Waxy genes role in starch formation, comparative analysis of the two data sets was undertaken (Fig. 3). There are three main starch granule morphologies within the cereals: (1) compound granulesdsmall rigidly polygonal granules arranged in large spherical compound granules typical of rice; (2) simple granulesdindividual large spherical/lenticellular to polygonal granules found in maize and sorghum; (3) bimodal granule arrangement with a mix of small spherical (B-type) and large lenticellular (A-type) granules typical of wheat (Tomlinson and Denyer, 2003). S. italica is the earliest diverging lineage within the PACCAD clade (Fig. 3). It is also the only member of this clade from this study with compound granules. Early SEM investigations of starch granule morphology of other members of this clade outside the Andropogoneae, revealed they too had compound granules within their endosperm, and it was proposed that this phenotype is the ancestral state (Shapter et al., 2008; Tateoka, 1962). This hypothesis would appear to be correct as both simple and bimodal granules occur only in the more recently diverged lineages. While the association between starch granule morphology and Waxy phylogeny appears to indicate a possible link between GBSSI and starch granule formation, it seems unlikely as it has previously been reported that a loss of function in this gene causes no apparent change to starch granule appearance (Buleon et al., 1998; Kim et al., 2003). Conversely, lesions in the isoamylase gene in barley were correlated to a lack of A and B-type granules and the appearance of compound granules in the endosperm (Burton et al., 2002). Screening of the amino acid alignment of all the species examined in this study observed several amino acid changes in the conserved domains of the gene which also correlated with the starch granule morphology (Fig. 2). These separated either all three types of granule (V), just the species with simple granules (U) or just the bimodal granules (x). While it is likely that these amino acid associations are due to the evolutionary history of the taxa, rather
than a direct genetic effect of the expression of a particular ortholog of GBSSI, it would be uncommon for an enzyme to have a single function (determination of amylose) with no other biochemical consequences. Perhaps as more is discovered about the role of GBSSI in endosperm development, these associations will be found to have more than just an evolutionary connection. Acknowledgements Seed for this study was supplied by Australian Tropical Crops and Forages Collection, Queensland Department of Primary Industries and Fisheries, www.dpi.qld.gov.au/auspgris/and Native Seeds Pty Ltd Australia http://www.nativeseeds.com.au. Funding was provided by the Grain Foods Cooperative Research Centre. References
Ashikari, M., Matsuoka, M., 2006. Identication, isolation and pyramiding of quantitative trait loci for rice breeding. Trends in Plant Science 11, 344350. Baldwin, P., 2001. Starch granule-associated proteins and polypeptides: a review. Starke. (Starch) 53, 475503. Bennetzen, J.L., Coleman, C., Liu, R.Y., Ma, J.X., Ramakrishna, W., 2004. Consistent over-estimation of gene number in complex plant genomes. Current Opinion in Plant Biology 7, 732736. Buleon, A., Colonna, P., Planchot, V., Ball, S., 1998. Starch granules: structure and biosynthesis. International Journal of Biological Macromolecules 23, 85112. Burton, R.A., Jenner, H., Carrangis, L., Fahy, B., Fincher, G.B., Hylton, C., Laurie, D.A., Parker, M., Waite, D., van Wegen, S., Verhoeven, T., Denyer, K., 2002. Starch granule initiation and growth are altered in barley mutants that lack isoamylase activity. The Plant Journal 31, 97112. Davis, J.P., Supatcharee, N., Khandelwal, R.L., Chibbar, R.N., 2003. Synthesis of novel starches in planta: opportunities and challenges. Starke. (Starch) 55, 107120. Demeke, T., Huel, P., Abdel-Aal, E.S.M., Baga, M., Chibbar, R.N., 1999. Biochemical characterization of the wheat waxy A protein and its effect on starch properties. Cereal Chemistry 76, 694698. Denyer, K., Johnson, P., Zeeman, S., Smith, A.M., 2001. The control of amylose synthesis. Journal of Plant Physiology 158, 479487. Domon, E., Saito, A., Takeda, K., 2002. Comparison of the waxy locus sequence from a non-waxy strain and two waxy mutants of spontaneous and articial origins in barley. Genes & Genetic Systems 77, 351359. Evans, R.C., Campbell, C.S., 2002. The origin of the apple subfamily (Maloideae; Rosaceae) is claried by DNA sequence data from duplicated GBSSI genes. American Journal of Botany 89, 14781484. Feuillet, C., Keller, B., 2002. Comparative genomics in the grass family: molecular characterization of grass genome structure and evolution. Annals of Botany 89, 310. Fortune, P.M., Schierenbeck, K.A., Ainouche, A.K., Jacquemin, J., Wendel, J.F., Ainouche, M.L., 2007. Evolutionary dynamics of Waxy and the origin of hexaploid Spartina species (Poaceae). Molecular Phylogenetics and Evolution 43, 10401055. Gaines, C.S., Raeker, M.O., Tilley, M., Finney, P.L., Wilson, J.D., Betchel, D.B., Martin, R.J., Seib, P.A., Lookhart, G.L., Donelson, T., 2000. Associations of starch gel hardness, granule size, waxy allelic expression, thermal pasting, milling quality, and kernel texture of 12 soft wheat cultivars. Cereal Chemistry 77, 163186. Hajjar, R., Hodgkin, T., 2007. The use of wild relatives in crop improvement: a survey of developments over the last 20 years. Euphytica 156, 113. Jansen, R., Ledley, F.D., 1990. Disruption of phase during PCR amplication and cloning of heterozygous target sequences. Nucleic Acids Research 18, 51535156. Jensen, M., Straus, N., 1993. Effect of PCR conditions on the formation of heteroduplex and single-stranded DNA products in the amplication of bacterial ribosomal DNA spacer regions. PCR Methods Application 3, 186194. Jozwik, F.X., 1969. Some systematic aspects of Mitchell grasses. Australian Journal of Botany 17, 359374. Kawase, M., Fukunaga, K., Kato, K., 2005. Diverse origins of waxy foxtail millet crops in east and southeast Asia mediated by multiple transposable element insertions. Molecular Genetics and Genomics 274, 131140. Kellogg, E.A., 2002. Classication of the grass family. In: Mallett, K., Orchard, A. (Eds.), Flora of Australia. ABRS/CSIRO Australia, Melbourne, pp. 1936. Khush, G.S., 2001. Challenges for meeting the global food and nutrient needs in the new millennium. Proceedings of the Nutrition Society 60, 1526. Kim, W., Johnson, J.W., Graybosch, R.A., Gaines, C.S., 2003. Physicochemical properties and end-use quality of wheat starch as a function of waxy protein alleles. Journal of Cereal Science 37, 195204. Kumar, A., Bennetzen, J., 1999. Plant retrotransposons. Annual Reviews in Genetics 33, 479532. Langridge, P., Lagudah, E.S., Holton, T.A., Appels, R., Sharp, P.J., Chalmers, K.J., 2001. Trends in genetic and genome analyses in wheat: a review. Australian Journal of Agricultural Research 52, 10431077. Longeri, M., Zanotti, M., Damiani, G., 2002. Recombinant DRB sequences produced by mismatch repair of heteroduplexes during cloning in Escherichia coli. European Journal of Immunogenetics 29, 517523.
F.M. Shapter et al. / Journal of Cereal Science 49 (2009) 411 Mangalika, W.H.A., Miura, H., Yamauchi, H., Noda, T., 2003. Properties of starches from near-isogenic wheat lines with different Wx protein deciencies. Cereal Chemistry 80, 662666. Marchler-Bauer, A., Anderson, J.B., Derbyshire, M.K., DeWeese-Scott, C., Gonzales, N.R., Gwadz, M., Hao, L., He, S., Hurwitz, D.I., Jackson, J.D., Ke, Z., Krylov, D., Lanczycki, C.J., Liebert, C.A., Liu, C., Lu, F., Lu, S., Marchler, G.H., Mullokandov, M., Song, J.S., Thanki, N., Yamashita, R.A., Yin, J.J., Zhang, D., Bryant, S.H., 2007. CDD: a conserved domain database for interactive domain family analysis. Nucleic Acids Research 35, D237D240. Mason-Gamer, R.J., Weil, C.F., Kellogg, E.A., 1998. Granule-bound starch synthase: structure, function, and phylogenetic utility. Molecular Biology and Evolution 15, 16581673. McIntosh, S.R., Pacey-Millar, T., Henry, R.J., 2005. A universal protocol for identication of cereals. Journal of Cereal Science 41, 3746. Murai, J., Taira, T., Ohta, D., 1999. Isolation and characterization of the three Waxy genes encoding the granule-bound starch synthase in hexaploid wheat. Gene 234, 7179. Nakamura, T., Yamamori, M., Hirano, H., Hidaka, S., Nagamine, T., 1995. Production of waxy (amylose-free) wheats. Molecular and General Genetics 248, 253259. Ochieng, J.W., Muigai, A.W.T., Ude, G.N., 2007. Phylogenetics in plant biotechnology: principles, obstacles and opportunities for the resource poor. African Journal of Biotechnology 6, 639649. Pedersen, J.F., Bean, S.R., Funnell, D.L., Graybosch, R.A., 2004. Rapid iodine staining techniques for identifying the waxy phenotype in sorghum grain and waxy genotype in sorghum pollen. Crop Science 44, 764767. Pedersen, J.F., Graybosch, R.A., Funnell, D.L., 2007. Occurrence of the waxy alleles wxa and wxb in waxy sorghum plant introductions and their effect on starch thermal properties. Crop Science 37. Preston, J.C., Kellogg, E.A., 2006. Reconstructing the evolutionary history of paralogous APETALA1/FRUITFULL-like genes in grasses (Poaceae). Genetics 174, 421437. Rahman, S., Li, Z., Batey, I., Cochrane, M.P., Appels, R., Morell, M., 2000. Genetic alteration of starch functionality in wheat. Journal of Cereal Science 31, 91110. Ramakrishna, W., Ma, J.X., SanMiguel, P., Emberton, J., Dubcovsky, J., Shiloff, B.A., Jiang, Z.Y., Rostoks, N., Busso, C.S., Ogden, M., Linton, E., Kleinhofs, A., Devos, K.M., Messing, J., Bennetzen, J.L., 2002. Frequent genic rearrangements in
11
two regions of grass genomes identied by comparative sequence analysis. Comparative and Functional Genomics 3, 165166. Rao, N.K., Reddy, L.J., Bramel, P.J., 2003. Potential of wild species for genetic enhancement of some semi-arid food crops. Genetic Resources and Crop Evolution 50, 707721. Shapter, F.M., Lee, L.S., Henry, R.J., 2008. Endosperm and starch granule morphology in wild cereal relatives. Plant Genetic Resources 6, 8597. Shure, M., Wessler, S., Fedoroff, N., 1983. Molecular identication and isolation of the Waxy locus in maize. Cell 35, 225233. Small, R.L., Cronn, R.C., Wendel, F., 2004. L.A.S. Johnson Review No. 2: use of nuclear genes for phylogeny reconstruction in plants. Australian Systematic Botany 17, 145170. Swofford, D.L., 2001. PAUP: Phylogenetic Analysis Using Parsimony (*and Other Methods), version 4.0b10. Sinauer Associates, USA. Taira, T., Fujita, N., Takaoka, K., Uematsu, M., Wadano, A., Kozaki, S., Okabe, S., 1995. Variation in the primary structure of waxy proteins (granule-bound starch synthase) in diploid cereals. Biochemical genetics 33, 269281. Tateoka, T., 1962. Starch grains of endosperm in grass systematics. Botany Magazine Tokyo 75, 377383. Thompson, J.R., Marcelino, L.A., Polz, M.F., 2002. Heteroduplexes in mixed-template amplications: formation, consequence and elimination by reconditioning PCR. Nucleic Acids Research 30, 20832088. Tomlinson, K., Denyer, K., 2003. Starch synthesis in cereal grains. Advances in Botanical Research 40, 161. Vaughan, D.A., Morishma, H., Kadowaki, K., 2003. Diversity in the Oryza genus. Current Opinion in Plant Biology 6, 139146. Vrinten, P.L., Nakamura, T., 2000. Wheat granule-bound starch synthase l and ll are encoded by separate genes that are expressed in different tissues. Plant Physiology 122, 255263. Wang, Z.Y., Zheng, F.Q., Gao, J.P., Wang, X.Q., Wu, M., Zhang, J.L., Hong, M.M., 1994. Identication of two transposon-like elements in rice Wx gene. Science in China 37, 437447. Williams, W., Verry, I.M., Ellison, N.W., 2006. A phylogenetic approach to germplasm use in clover breeding. In: Woodeld, D. (Ed.), 13th Australasian Plant Breeders Conference, Christchurch, NZ. Yan, L., Bhave, M., Fairclough, R., Konik, C., Rahman, S., Appels, R., 2000. The genes encoding granule-bound starch synthase at the waxy loci of the A, B and D progenitors of common wheat. Genome 43, 264272.

Effect of high temperature on albumin and globulin accumulation in the endosperm proteome of the developing wheat grainq
William J. Hurkman*, William H. Vensel, Charlene K. Tanaka, Linda Whitehand, Susan B. Altenbach
U.S. Department of Agriculture, Agricultural Research Service, Western Regional Research Center, 800 Buchanan Street, Albany, CA 94710, USA
Article history: Received 14 December 2007 Received in revised form 6 June 2008 Accepted 13 June 2008 Keywords: Albumins Globulins Grain ll High temperature
a b s t r a c t
The accumulation of KCl-soluble/methanol-insoluble albumins and globulins was investigated in the endosperm of developing wheat (Triticum aestivum, L. cv. Butte 86) grain produced under a moderate (24 C/17 C, day/night) or a high temperature regimen (37 C/28 C) imposed from 10 or 20 days postanthesis (dpa) until maturity. Proteins were separated by 2-DE and developmental proles for nearly 200 proteins were analyzed by hierarchical clustering. Comparison of protein proles across physiologically equivalent stages of grain ll revealed that high temperature shortened, but did not substantially alter, the developmental program. Accumulation of proteins shifted from those active in biosynthesis and metabolism to those with roles in storage and protection against biotic and abiotic stresses. Few proteins responded transiently when plants were transferred to the high temperature regimens, but levels of a number of proteins were altered during late stages of grain development. Specic protein responses depended on whether the high temperature regimens were initiated early or mid development. Some of the heat responsive proteins have been implicated in gas bubble stabilization in bread dough and others are suspected food allergens. Published by Elsevier Ltd.
1. Introduction High temperature during grain ll is one of the more signicant environmental factors that affects wheat yield and our quality (Skylas et al., 2002 and references therein). Although major processes in grain ll have been described (Laudencia-Chingcuanco et al., 2007; McIntosh et al., 2007; Vensel et al., 2005), little is known about the effects of high temperature on this developmental program. High temperature shortens the duration of grain ll and decreases the time to apoptosis and harvest maturity (Altenbach et al., 2003). Consistent with these events, transcripts for a-, g-, and u-gliadins, high molecular weight glutenin subunits (HMW-GS), and low molecular weight glutenin subunits (LMW-GS) accumulate and disappear earlier (Altenbach et al., 2002; Altenbach and Kothari, 2004). Transcripts for genes functioning in protein synthesis, starch synthesis, stress/defense, as well as storage accumulate earlier in response to high temperature (Altenbach and
Abbreviations: dpa, days post-anthesis; HMW-GS, high molecular weight glutenin subunits; LMW-GS, low molecular weight glutenin subunits. q Disclaimer: The mention of a trademark or proprietary product does not constitute a guarantee or warranty of the product by the United States Department of Agriculture and does not imply its approval to the exclusion of other products that may be suitable. * Corresponding author. Tel.: 1 510 559 5720; fax: 1 510 559 5818. E-mail address: william.hurkman@ars.usda.gov (W.J. Hurkman). 0733-5210/$ see front matter Published by Elsevier Ltd. doi:10.1016/j.jcs.2008.06.014
Kothari, 2004; Hurkman et al., 2003). High temperatures during grain ll inuence gluten protein accumulation levels. The relative amounts of certain a-gliadins and HMW-GS were higher and those for certain LMW-GS were lower in grain produced under a 37 C/ 28 C day/night regimen (DuPont et al., 2006b). In addition, accumulation rates increased more for the a-gliadins and HMW-GS than the LMW-GS in grain produced under this regimen (DuPont et al., 2006a). High temperature during grain ll also affects the accumulation levels of stress/defense proteins. Skylas et al. (2002) reported that heat shock, a shift from a 24 C/18 C to a 40 C/25 C day/night regimen from 1517 dpa, increased the number of small heat shock protein (HSP) isoforms in total protein extracts from endosperm. Majoul et al. (2003) reported that levels of HSPs as well as proteins that defend against reactive oxygen species (ROS) and desiccation increased in our when developing grain was exposed to an extended high temperature day/night regimen of 34 C/10 C from anthesis to maturity. Majoul et al. (2004) isolated an albumin and globulin fraction from mature grain and found that the levels of a number of proteins, including HSPs as well as enzymes involved in starch biosynthesis, carbohydrate metabolism, and ATP synthesis, responded to the high temperature regimen. In this study, we utilize proteome maps to develop a comprehensive picture of endosperm development and identify the effects of high temperature on protein composition during this crucial reproductive stage. Plants were grown under three temperature conditions a moderate temperature regimen maintained
W.J. Hurkman et al. / Journal of Cereal Science 49 (2009) 1223
13
throughout grain development and a high temperature regimen initiated early or mid development and maintained until maturity. The high temperature regimens are more severe than might be encountered under eld conditions, but were selected to accentuate protein responses. KCl-soluble/methanol-insoluble albumins and globulins were isolated from the endosperm at selected time points during grain development and separated by 2-DE. These proteins function in a wide range of cellular processes (Vensel et al., 2005), including the synthesis of gluten proteins and starch, major determinants of wheat yield and grain quality. Protein accumulation proles and functions were compared across physiologically equivalent stages of grain ll to dene major events of endosperm development and identify specic responses to high temperature. Protein levels in mature grain produced under the three temperature regimens were also compared to discover changes in protein composition related to grain quality. 2. Materials and methods 2.1. Plant material Triticum aestivum, L. cv. Butte 86, was grown in climatecontrolled greenhouses under moderate and high temperature regimens (Altenbach et al., 2007b). For the moderate temperature regimen, plants were grown at a maximum daytime temperature of 24 C and a minimum nighttime temperature of 17 C. For the high temperature regimen, plants were transferred at 10 or 20 days postanthesis (dpa) to a second greenhouse and grown at a maximum daytime temperature of 37 C and a minimum nighttime temperature of 28 C. Water and fertilizer (Plantex 20-20-20, 300 mg/day) were applied by drip irrigation. Natural light was supplemented with 100 W high-pressure sodium lamps to maintain a day length of 16 h. Heads were collected at selected time points during grain development. Under the moderate temperature regimen, heads were collected at 4-day intervals from 10 to 38 dpa and at 40 dpa. Because high temperature shortened the grain developmental program, heads were collected at 2-day intervals from 10 to 26 dpa for the high temperature regimen initiated at 10 dpa and from 20 to 28 dpa for the high temperature regimen initiated at 20 dpa. The end point for each regimen was the oldest age that endosperm could be squeezed or scraped from the grain. Grains were removed from the heads and the region containing the embryo was excised with a razor blade. The endosperm was squeezed through the resultant opening in the pericarp/testa. Endosperm was transferred immediately into tubes cooled in liquid nitrogen and stored at 80 C. Endosperm was also collected from plants grown in three additional experiments in which a 37 C/28 C or a 37 C/17 C high temperature regimen with 300 mg per day fertilizer was initiated at anthesis and a 37 C/28 C high temperature regimen with 150 mg per day fertilizer was initiated at 15 dpa. 2.2. 2-DE analysis A KCl-soluble/methanol-insoluble albumin and globulin fraction was prepared from the endosperm and triplicate 2-D gels loaded with equal protein amounts from each time point were run as described previously (Hurkman and Tanaka, 2004; Vensel et al., 2005). Gels were digitized with a calibrated scanner (UMAX Powerlook III; Dallas, TX) at 300 dpi using the same settings for all gels. Computer software (Progenesis PG240 Ver. 2006; Nonlinear Dynamics Limited, Newcastle upon Tyne, UK) was used to match protein patterns and analyze the quantitative and qualitative differences across time points for the three temperature regimens; matching was validated manually for all spots. Protein accumulation proles were constructed from the normalized spot volume (individual spot volume/total spot volume 100) data and validated manually.
2.3. Protein identication Proteins contained in 2-D gel spots were digested with trypsin using a DigestPro gel-spot-processing robot (Intavis, Lagenfeld, Germany). Peptides from the tryptic digests were identied using a QSTAR PULSAR i quadrupole time-of-ight (TOF) mass spectrometer (Applied Biosystems/MDS SCIEX, Toronto, Canada) as previously described (Vensel et al., 2005). The resulting AnalystQS wiff les were converted to MGF les using Mascot Daemon (http:// www.matrixscience.com/) and data analysis of the MS/MS les was carried out using a locally installed copy of the Global Proteome Machine Organization (GPM) software (http://www.thegpm.org). X!Tandem (version 2005.06.01.2), the spectrum modeler in the GPM software, was used to match MS/MS fragmentation data to peptide sequences. MGF les were submitted in batch mode to X!Tandem using a script provided by Jayson Faulkner (University of Michigan). X!Tandem was congured to search a le containing 695,000 protein sequence entries from all proteins in the HarvEST:Wheat version 1.04 database (http://harvest.ucr.edu/ HWheat104.exe), NCBI nonredundant green plant database, NCBI Triticum aestivum: UniGene Build #37, and wEST Database (http:// wheat.pw.usda.gov/wEST). A local installation of the GPM and the GPMDB (Craig et al., 2004) was used to analyze, store, and display the data. The protein identication and mass spectrometry data are available as XML les (http://wheat.pw.usda.gov/pubs/2008). To access the data, click on Hurkman, scroll down to Compressed directory of individual XML les for gel spots, click to download, and then unzip the les. To view the les, access the GPM viewer (http://h777.thegpm.org/tandem/thegpm_upview.html), click Browse, select the le for the spot number of interest, and click View Models. Click Protein to see the protein sequence and identied peptides. 2.4. Statistical analyses The protein accumulation data was analyzed by hierarchical cluster analysis to determine similarities and differences across the developmental time courses for the moderate and high temperature regimens. The data set consisted of accumulation proles for 193 proteins. Each prole consisted of 23 time points across the three temperature regimens. In order to discover patterns independent of protein abundance, the data for each protein were converted to a percent scale (normalized volume of a given time point/total normalized volume for the 23 time points 100). Hierarchical cluster analysis was performed with SAS software, PROC CLUSTER, using the average method. Statistical algorithms, including CCC, pseudo t, and pseudo F were used to choose the nal number of clusters (SAS Institute Inc., 2004). To determine reproducibility of the data for effects of high temperature on endosperm protein levels late in grain development, correlations were computed for the data from the high temperature regimens initiated at anthesis, 10, 15, and 20 dpa. To ensure that the correlations were not unduly inuenced by large observations, the Kendall correlation method, based on ranks across all proteins, was performed using the SAS software PROC CORR (SAS Institute Inc., 2004). In addition Cronbachs alpha was computed to estimate consistency among regimens. 3. Results 3.1. Endosperm proteome maps In a previous study (Vensel et al., 2005), we developed proteome maps for albumins and globulins isolated from the endosperm of grain grown under moderate temperatures (24 C/17 C, day/ night). The maps, which contain 254 proteins, were established
14
using two developmental time points, 10 and 36 dpa, to maximize proteome coverage. In the present study, we mapped and identied 35 additional endosperm proteins (Fig. S1). Spot numbers, accession numbers, percent coverage, expectation values, and peptide sequence data for these proteins are listed in Table S1. Of the 289 proteins in the updated proteome maps, 193 were selected for prole analysis. Spots that contained more than one protein, 90 in all, were eliminated from the experimental set, because normalized volumes for the individual proteins within these spots could not be determined. Since this study focuses on the albumins and globulins, the 6 spots containing gliadins and glutenins that co-puried with this fraction (Vensel et al., 2005) also were eliminated from the data set. 3.2. Hierarchical cluster analysis Accumulation proles based on normalized spot volumes were established for proteins in endosperm harvested from grain grown under the moderate and high temperature regimens (Fig. 1). The data set for the high temperature regimen initiated at 20 dpa includes spot volumes from the three initial time points (10, 14, and 18 dpa) of the moderate temperature regimen under which the plants were grown before transfer to this high temperature regimen. When normalized protein volumes were summed for all time points across the three temperature regimens for each protein, the totals spanned four orders of magnitude, ranging from 0.13 for an isoform of globulin-2 (#561) to 118 for protein disulde isomerase (PDI, #118). Because of this wide dynamic range, the normalized volume data were converted to percentages so that hierarchical cluster analysis results would be independent of protein amount. In Fig. 1, the protein prole data were converted to colors: white represents the lowest percentage range (00.75), shades of blue the intermediate percentage ranges (>0.7540), and black the highest percentage range (>40). The identities of the proteins corresponding to the proles depicted in Fig. 1 and their functions are listed in Table 1 by the order and cluster numbers determined by the hierarchical cluster analysis (Fig. S2). The order numbers in Table 1 may be used to identify proteins shown in Fig. 1. The spot numbers in Table 1 may be used to locate the spots on the proteome maps in Fig. S1 of this paper and Figs. 1 and 2 of Vensel et al. (2005). The hierarchical cluster analysis separated the 193 protein accumulation proles into 39 clusters (Fig. 1 and Fig. S2), but nearly 65% of the proteins in the analysis were members of four clusters (1, 2, 3, and 5). Cluster 1 with 58 members was the largest cluster. The majority of these proteins functioned in carbohydrate metabolism, protein synthesis/assembly, and nitrogen metabolism. Cluster 5 with 40 members was the second largest cluster. The majority of the proteins in this cluster functioned in stress/defense, although some also functioned in carbohydrate and nitrogen metabolism. The next largest clusters were 2 and 3 with 16 and 13 members, respectively. In contrast to cluster 1, the majority of the proteins in these clusters functioned in transcription/translation. The remaining clusters, 4 and 639, were smaller, containing 110 members, and the majority of these proteins functioned in carbohydrate metabolism, protein synthesis/assembly, storage and stress/defense. The most abundant proteins (total normalized volume >15, indicated by purple boxes in Fig. 1 and an H in Table 1) were members of the four largest clusters (1, 2, 3, and 5). The least abundant proteins (total normalized volume <1, indicated by dark pink boxes in Fig. 1 and an L in Table 1) were found in many different clusters. Examination of the protein proles in Fig. 1 reveals that they are ordered by the timing of peak accumulation levels during the developmental time course. Proteins in clusters 13 peak early in development, proteins in cluster 5 peak during mid development, and proteins in the majority of clusters 636 peak late in development.
3.3. Effect of the high temperature regimens on the grain developmental program Fresh weight accumulation proles (Fig. 2) were utilized to determine physiologically equivalent stages of grain ll for the three temperature regimens. The high temperature regimens shortened the duration of grain ll. Grain matured by 44 dpa under the moderate temperature regimen compared to 28 and 33 dpa under the high temperature regimens initiated at 10 and 20 dpa, respectively (Fig. 2). In addition, maximal fresh weight/kernel was attained at 38 dpa under the moderate temperature regimen compared to 22 and 24 dpa under the high temperature regimens initiated at 10 and 20 dpa, respectively. Maximal fresh weight/kernel was reduced from 82 mg under the moderate temperature regimen to 53 and 65 mg under the high temperature regimens initiated at 10 and 20 dpa, respectively. Final kernel weight was also reduced by the high temperature conditions. Under the moderate temperature regimen, nal weight/ kernel was 69 mg compared to 29 and 38 mg under the high temperature regimens initiated at 10 and 20 dpa, respectively. Based on the fresh weight accumulation proles (Fig. 2), grain development was divided into three stages. The early stage began at 10 dpa and ended when the kernel attained one-half of maximum fresh weight, the middle stage ended when the kernel attained maximal fresh weight, and the late stage ended at kernel maturity. Under the staging system for small grains developed by Zadoks et al. (1974), these stages correspond to late milk (start of grain ll), soft dough (maximum fresh weight), and hard kernel (ripening) stages of grain development. Under the moderate temperature regimen, the early, middle, and late stages included the 1018, 2234, and 3840 dpa time points, respectively. Under the high temperature regimen initiated at 10 dpa, the early, middle, and late developmental stages included the 1014, 1622, and 2426 dpa time points and under the high temperature regimen initiated at 20 dpa, they included the 1018, 2024, and 2628 dpa time points. The number of proteins that peaked during the early, middle, and late stages was determined for the three temperature regimens (Table 1) and graphed with respect to function (Fig. 3). This analysis showed a shift in the accumulation of proteins during grain ll from those active in biosynthesis and metabolism to those with roles in storage and protection against biotic and abiotic stresses. Under the moderate temperature regimen, proteins that peaked early in development functioned principally in carbohydrate metabolism, protein synthesis/assembly, stress/defense and transcription/ translation; nitrogen metabolism and protein turnover were also important functions at this developmental stage (Fig. 3A). During mid development, the number of proteins functioning in stress/ defense remained at similar levels while the number of those involved in carbohydrate metabolism, nitrogen metabolism, protein synthesis/assembly, protein turnover, signal transduction, and transcription/translation decreased substantially (Fig. 3B). Late in development, stress/defense and storage were major protein functions followed distantly by carbohydrate metabolism and protein synthesis, while proteins functional in nitrogen metabolism were not detected (Fig. 3C). Under the high temperature regimens initiated at 10 and 20 dpa, the distribution of proteins in each functional category was similar during early, mid, and late developmental stages to that for the moderate temperature regimen (Fig. 3AC). In fact, 87% of the proteins associated with the early, mid and late stages of the high temperature regimen initiated at 10 dpa corresponded to those associated with the early, mid, and late stages of the moderate temperature regimen. Similarly, 89% of the proteins associated with the early, mid and late stages of the high temperature regimen initiated at 20 dpa corresponded to those associated with the early, mid, and late stages of the moderate temperature regimen. Despite the shortened duration of grain ll with the accompanying decrease in kernel weight, this
15
Fig. 1. Accumulation patterns of KCl-soluble/methanol-insoluble albumins and globulins in wheat endosperm. Proteins were isolated from grain produced under moderate (24 C/ 17 C) and high temperature conditions (37 C/28 C from 10 or 20 dpa). Proles are shown as percent of total normalized volume. Hierarchical clusters (139) and accompanying protein order were determined by SAS analysis (see Fig. S1). Protein identications are in Table 1. The most abundant and least abundant proteins, based on total normalized spot volume for the three treatments, are indicated in the Abundance column. Differences in proles between the 24 C/17 C and the 37 C/28 C from 10 or 20 dpa regimens are indicated in the Changes with Heat columns.
16
Table 1 Identities of endosperm proteins corresponding to the accumulation proles shown in Fig. 1 Order no. Cluster no. Spot no. Name Functiona Peak timingb 24/17 37/28 10 dpa 37/28 20 dpa 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 264 294 262 112 252 383 97 118 176 138 154 250 354 6 69 72 159 433 65 145 103 104 96 173 135 295 48 301 415 444 157 158 411 422 153 655 98 25 27 179 185 161 82 117 1004 373 278 306 163 74 149 285 61 378 447 49 471 342 344 119 193 657 70 178 53 66 57 60 63 31 55 493 529 361 Glyceraldehyde 3-P dehydrogenase (NAD) Malate dehydrogenase (NAD) Glyceraldehyde 3-P dehydrogenase (NAD) PPi-fructose 6-P 1-phosphotransferase b-subunit Aldolase Cyclophilin A-2 Fructokinase Protein disulde isomerase Alanine amino transferase 2 Ketol-reductoisomerase Aldehyde dehydrogenase Reversibly glycosylated polypeptide Triosephosphate isomerase Carbamoyl phosphate synthetase Poly(A)-binding protein 10-Formyltetrahydrofolate synthetase ADP-glucose PPase, SS Gycine-rich RNA-binding protein Heat shock protein 70 Selenium binding protein Phosphoglycerate mutase, 2,3-bisphosphoglycerate-independent Phosphoglycerate mutase, 2,3-bisphosphoglycerate-independent Phosphoglucomutase Alanine amino transferase 2 Leucine amino peptidase Malate dehydrogenase (NAD) Methionine synthase Guanine nucleotide-binding protein b subunit-like protein Ubiquitin-protein ligase 40S Ribosomal protein S21 UDP-glucose PPase UDP-glucose PPPase Gycine-rich RNA-binding protein 40S Ribosomal protein S12 Aldehyde dehydrogenase S-Adenosylmethionine synthetase 2 Acetohydroxyacid synthase Aconitase Aconitase Enolase Enolase Dihydrolipoamide dehydrogenase Stress-induced protein, sti1-like Chaperonin 60 kDa b-subunit Fructose bisphosphate aldolase 20S Proteasome a-subunit B TGF-b receptor-interacting protein 1 Aldolase Heat shock associated protein DNAK-type molecular chaperone HSP70 Heat shock associated protein Legumin-like protein Poly(A)-binding protein GSH-dependent dehydroascorbate reductase 1 Polyubiquitin 6 Heat shock protein 80-2 DNAK-type molecular chaperone HSP70 Ascorbate peroxidase Ascorbate peroxidase Phosphoglycerate dehydrogenase-like protein 6-Phosphogluconate dehydrogenase S-Adenosylmethionine synthetase 1 Heat shock protein 70 Enolase Poly(A)-binding protein Poly(A)-binding protein Poly(A)-binding protein Poly(A)-binding protein Poly(A)-binding protein Pyruvate Pi dikinase Poly(A)-binding protein UDP-glucose dehydrogenase 40S Ribosomal protein S19 20S Proteasome a-subunit CM CM CM CM CM PS CM PS NM NM CM CM CM NM TT NM CM TT PS S CM CM CM NM PT CM NM ST PT PS CM CM TT PS S NM NM CM CM CM CM CM S PS CM PT ST CM PS PS PS SP TT S PT PS PT S S CM CM NM PS CM TT TT TT TT TT CM TT PS PS PT E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E H H H H H H H Abundancec
L L L L
W.J. Hurkman et al. / Journal of Cereal Science 49 (2009) 1223 Table 1 (continued ) Order no. Cluster no. Spot no. Name Functiona Peak timingb 24/17 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 3 3 3 3 3 3 3 3 3 3 3 3 4 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 6 6 6 6 6 7 8 9 9 9 9 9 9 9 9 9 9 10 11 11 12 13 399 362 94 182 412 201 667 348 330 640 279 93 42 350 382 356 110 205 232 175 353 417 43 379 430 111 436 384 440 381 386 245 369 434 233 236 231 246 237 33 305 128 132 130 148 244 251 249 206 312 137 418 102 141 337 336 327 335 1009 199 255 259 263 371 314 324 168 212 441 639 407 225 338 666 341 Translationally controlled tumor protein Elongation factor 1-b RNA-binding protein, similarity Tubulin a-3 chain Glycine-rich RNA-binding protein SGT1 Ascorbate peroxidase Ascorbate peroxidase 14-3-3 protein Phosphoglycerate kinase Adenosine kinase RNA-binding protein, similarity Heat shock protein, 82 K Small ras-related GTP-binding protein Dehydroascorbate reductase Triosephosphate isomerase 2-Isopropylmalate synthase 26S Proteasome regulatory particle triple-A ATPase subunit 4 Aspartate amino transferase Globulin-2 Avenin a-Amylase/trypsin inhibitor, CM3 Late embryogenesis abundant protein-like Expressed protein a-Amylase inhibitor 0.19 PPi-fructose-6-P 1-phosphotransferase a-Amylase inhibitor 0.53 Avenin N9 a-Amylase inhibitor Ima1, monomeric a-Amylase/subtilisin inhibitor F10K1.21/F7A7_100 protein, similarity Serpin WZS3 Avenin N9 a-Amylase inhibitor 0.53 Serpin Aspartate amino transferase Serpin WZS2 Serpin Serpin WZS3 Elongation factor 2 Glyoxalase I ADP-glucose PPase, LS ADP-glucose PPase, LS ADP-glucose PPase, LS Catalase isozyme 1 Aldolase Formate dehydrogenase Formate dehydrogenase Eukaryotic initiation factor 4A Purple acid phosphatase Leucine amino peptidase Superoxide dismutase [CuZn] Protein disulde isomerase-like protein Elongation factor 1-a Seed globulin Seed globulin Seed globulin Seed globulin Peroxidase 1 Triticin Peroxidase 1 Peroxidase 1 Peroxidase BP1 Peroxiredoxin OSJNBb0118P14.5 OSJNBb0118P14.5 Dihydrolipoamide dehydrogenase DNAK-type molecular chaperone HSP70 Avenin N9 Serpin Heat shock protein 16.9C Heat shock protein 70 DNA-binding protein HEXBP Ascorbate peroxidase Chitinase-c TT PS TT CD TT S S S ST CM A TT PS T S CM NM PT NM SP SP S S U S CM S SP S S U S SP S S NM S S S PS S CM CM CM S CM NM NM TT S PT S PS PS SP SP SP SP S SP S S S S U U CM PS SP S PS PS TT S S E E E E E E E E E E E E E M M M M M M M M M M M M M M M M M M M E M E E E L L L M M M M M M M M M L L M L L L L L L L M L L L L L L L L L L M E E E L 37/28 10 dpa 37/28 20 dpa E E E E E E E E E E E E E M M M M M M M M M L M M M M L M L L L M M E E E M M M L M M M M M M M E E L L M L M M M M M L L L L L L L L L L L M E E E L E E E E E E E E E E E E E M M M M M M M L L L M M M M M M L L M E L M M M M M L L M M M M M M M M M L L L L L L M M M M L L L L L L L L L L M E E E L
17
Abundancec
H H L H
H H H
H H
(continued on next page)
18 Table 1 (continued ) Order no. Cluster no. Spot no. Name
Functiona Peak timingb 24/17 37/28 10 dpa 37/28 20 dpa
Abundancec
150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193
13 13 13 13 13 13 13 14 14 14 15 15 15 15 15 15 16 17 18 18 19 20 20 21 22 23 24 25 26 27 28 29 30 31 31 32 33 34 35 36 36 37 38 39
351 282 842 1001 557 551 427 318 550 633 355 591 275 347 420 281 147 376 155 620 631 1012 1014 310 283 359 1010 419 672 357 869 125 1007 559 851 852 871 561 1022 1002 1003 311 530 364
Chitinase-a S Chitinase-a S Late embryogenesis abundant protein-like S Unknown protein [similar to late embryogenesis abundant protein-like] S Embryo-specic protein U Late embryogenesis abundant protein-like S Pathogenesis-related protein 4 S Xylanase inhibitor I S Sucrose synthase type 2 CM Nonspecic lipid-transfer protein T Seed globulin SP Glucose and ribitol dehydrogenase CM Glyceraldehyde 3-P dehydrogenase (NAD) CM PT 20S Proteasome a-subunit Nucleoside diphosphate kinase I A Globulin Beg 1 SP Globulin-2 SP Superoxide dismutase [Mn] S Globulin-2 SP Globulin Beg 1 SP Globulin-like protein SP Peroxidase 1 S Peroxidase 1 S Glucose and ribitol dehydrogenase CM ST TGF-b receptor-interacting protein 1 Peroxidase 1 S Peroxidase 1 S Glycine-rich RNA-binding protein TT Reversibly glycosylated polypeptide CM Triosephosphate isomerase CM 60S Ribosomal protein L12 PS b-Amylase CM Glyceraldehyde 3-P dehydrogenase (NAD) CM Globulin-2 SP Globulin-2 SP Globulin-2 SP Barwin S Globulin-2 SP Embryo globulin SP Embryo globulin SP Embryo globulin SP Initiation factor 3g TT Cyclophilin A-1 PS Ascorbate peroxidase S
L L L L L L L L L L L L L L L E L L L L L L L L E E M M E E L L L L L L L L A L L E E E
L L L L L L L L L L L L L L L L L L L L L A A M A M L E E E L L A L L L L L L L L E E E
L L L L L L L L L L L L L L L E L L L L L M M E E E M E E E L L L L L A L A A L L E E E
L L
L L L L L L L
Protein clusters were determined using SAS as outlined in Section 2. Spot numbers correspond to 2-DE map locations in Fig. S1 and Figs. 1 and 2 in Vensel et al. (2005). a A, ATP interconversion; CD, cell division; CM, carbohydrate metabolism; NM, nitrogen metabolism; PS, protein synthesis/assembly; PT, protein turnover; S, stress/defense; SP, storage protein; ST, signal transduction; T, transport; TT, transcription/translation; U, unknown. b E, early; M, middle; L, late stages. See text for time points included for each of the temperature regimens. c H, high and L, low abundance proteins.
analysis of physiologically equivalent stages shows that the high temperature regimens did not substantially alter the overall pattern of protein accumulation during grain development. 3.4. Transient effects of high temperature on protein proles during grain development Developmental proles were examined for transient responses following transfer of plants to the high temperature regimens initiated at 10 and 20 dpa. Proteins that increased (yellow boxes in Fig. 1) or decreased (purple boxes in Fig. 1) transiently in response to the high temperature regimens are listed in Table 2. Under the high temperature regimen initiated at 10 dpa, two HSP70 isoforms (225, 471), legumin-like protein (285), and triosephosphate isomerase (354) increased transiently early in development. HSP16.9 (407), four globulin isoforms (327, 335337), and a peroxidase (1009) increased while three isoforms of ADP-glucose PPase large subunit (128, 130, 132) decreased transiently during mid development under this regimen. Under the high temperature regimen initiated at 20 dpa, only two transient changes were detected.
HSP16.9 (407) increased and triticin (199) decreased transiently during mid development. Because the proteome maps were developed for endosperm proteins isolated from grain produced under the moderate temperature regimen, spots listed in Table 2 were excised from 2-D gels of endosperm proteins extracted from grain produced under the high temperature regimens and re-identied by MS/MS. The original protein identications were conrmed for all spots. The peptide sequence data, accession numbers, percent coverage, and expectation values are listed in Table S2. 3.5. Effects of high temperature regimens on endosperm protein levels late in grain development The high temperature regimens altered the accumulation levels of a number of proteins in the mature grain. Proteins that increased or decreased 1.8-fold or more relative to the moderate temperature regimen are listed in Table 3. Under the high temperature regimen initiated at 10 dpa, 31 proteins increased, the majority of which functioned in stress/defense and storage. The stress/defense
19
100
80
60
40
20
10
20
30
40
50
Days Post Anthesis (DPA)

Fig. 2. Effect of high temperature on grain fresh weight. Wheat plants were grown under three day/night temperature regimens: -, 24 C/17 C; C, 37 C/28 C initiated at 10 dpa; and B, 37 C/28 C initiated at 20 dpa.
proteins included a-amylase/subtilisin inhibitor (381), barwin/PR-4 protein (427, 871), chitinase (282, 341, 351), late embryogenesis abundant (LEA) protein (551, 1001), and xylanase inhibitor protein (318). It should be noted that lipid-transfer protein (LTP, 633), which is classied as a transport protein, may also have a role in stress/defense (Jung et al., 2003; Wu et al., 2004). Nearly all of the storage proteins that increased in response to high temperature were isoforms of globulin-2 (147, 155, 559, 561, 851, 852) or globulin Beg 1 (281, 620, 1002, 1003). In addition, 32 proteins decreased
40 30 20 10 0
under this high temperature regimen and the majority of these function in carbohydrate metabolism, including starch biosynthesis, and protein synthesis. Under the high temperature regimen initiated at 20 dpa, only 10 proteins increased and, like the regimen initiated at 10 dpa, the majority functioned in stress/defense and storage (Table 3). Eight of these proteins also increased during the high temperature regimen initiated at 10 dpa, but the increases in sucrose synthase type 2 (550), aspartate amino acid transferase (236), and CuZn superoxide dismutase (418) were specic for the regimen initiated at 20 dpa. In addition, 15 proteins decreased, including 7 that also decreased under the high temperature regimen initiated at 10 dpa. However, decreases in 40S ribosomal protein S12 (422) and LEA protein (842) were specic for the 20 dpa regimen. Six proteins, including ve globulins (147, 561, 620, 631, 852) and a 40S ribosomal protein S21 (444), that increased under the high temperature regimen initiated at 10 dpa, decreased under the 20 dpa regimen. Like the proteins that responded transiently to high temperature, proteins whose levels were altered late in development were also reidentied by MS/MS (Table S2). The original protein identications were conrmed for all but spot 69. This spot was initially identied as a poly(A)-binding protein, but was re-identied as formate tetrahydrofolate ligase. This is likely due to slight differences in the 2-DE pattern in this region of the gel, illustrating the importance of conrming identications of proteins in the high temperature proteomes. Nine of the re-identied spots contained peptides from one or two additional proteins (superscript a in Table 3). Because the proteins in these spots comigrate in the 2-D gels, it is not possible to know which proteins are responding to high temperature. Nonetheless, these identications are valuable because they provide additional candidates for heat responsive endosperm proteins. 3.6. Protein responses to high temperature regimens initiated at anthesis and 15 dpa Accumulation proles of albumins and globulins were analyzed in independent experiments in which grain was produced under a 24 C/17 C regimen and one of three additional high temperature regimens: 37 C/17 C initiated at anthesis, 37 C/28 C initiated at anthesis, or 37 C/28 C initiated at 15 dpa. Cronbachs alpha was computed for the control data from the last time point in these three experiments and the original experiment where heat was applied from 10 and 20 dpa. For this analysis, proteins that were absent at the last time point in one or more experiments were deleted, resulting in a data set of 163 proteins for which the calculated Cronbachs alpha was 0.96 out of a maximum of 1.0. This high level of consistency among the controls justied a comparison of the data obtained for the four high temperature experiments. At the last time point for each protein, the ratio of the normalized volume under the high temperature regimens to the moderate temperature regimen was calculated. Proteins that were absent at the last time point for one or more experiments were deleted, resulting in a data set containing 159 proteins, and the ratios were used to assign ranks to the proteins. Cronbachs alpha was computed for these ranks at the last time point for the four high temperature experiments. The calculated Cronbachs alpha of 0.60 showed a high degree of association among the four experiments. Individual correlations based on Kendalls Tau rank correlation method showed that the highest correlation was between the 37 C/28 C regimens initiated at anthesis and 10 dpa (0.31 at p < 0.0001). The 37 C/28 C regimen initiated at 10 dpa was also highly correlated with the 37 C/17 C regimen initiated at anthesis (0.18 at p < 0.0007), but less correlated with the 37 C/ 28 C regimens initiated at 15 (0.12 at p < 0.022) and 20 dpa (0.19 at p < 0.001). The 37 C/28 C regimens initiated at 15 and 20 dpa showed a high correlation with each other (0.22, signicant at
Average fresh weight/kernel (mg)
EARLY
Number of Proteins
MID
30 20 10 0 30 20 10 0
LATE
Ca
rb
.M
et
ab
. N M et
ab
. . ro Sy n ./A
ss
em
. o. T n ur ov
er St re ss /D
ef
en
se St or e ag
Pr
o. Si gn a
r lT
an
s. ns cr ip r ./T
an
sl.
Pr
Function
a Tr
Fig. 3. Effect of high temperature during grain ll on the timing of biochemical processes in wheat endosperm. See text for specic time points included in the early (A), mid (B), and late (C) stages for each of the three temperature regimens. ,, 24 C/ 17 C; , 37 C/28 C initiated at 10 dpa; and -, 37 C/28 C initiated at 20 dpa.
20
Table 2 Proteins that undergo transient changes in accumulation levels in response to high temperatures during grain ll Spot no. Name Functiona Order no.b Cluster no. Increasec 10 dpa 128 130 132 199 225 285 327 335 336 337 354 407 471 1009 ADP-glucose PPase, LS ADP0-glucose PPase, LS ADP-glucose PPase, LS Triticin Heat shock protein 70 Legumin-like protein Seed globulin Seed globulin Seed globulin Seed globulin Triosephosphate isomerase Heat shock protein 16.9C DNAK-type molecular chaperone HSP70 Peroxidase 1 CM CM CM SP PS SP SP SP SP SP CM PS PT S 116 118 117 134 146 52 131 132 130 129 13 145 57 133 5 5 5 8 11 1 9 6 6 6 1 10 1 7 20 dpa Decrease 10 dpa M M M M E E M M M M E M E M 20 dpa
M E
Identities of these proteins contained in these spots were conrmed by mass spectrometry. Accession numbers, percent coverage, expectation values, and peptide sequence data are listed in Table S2. XML les that contain the mass spectrometry data are available online (see Section 2.3). a Abbreviations for functional groups are: CM, carbohydrate metabolism; PS, protein synthesis/assembly; PT, protein turnover; S, stress/defense; SP, storage protein. b Order and cluster numbers were determined by hierarchal clustering. See Fig. 1 for accumulation proles. c E, transient change early in development; M, transient change in mid development. See text for specic time points.
p < 0.0001), lesser correlation with regimens initiated at 10 dpa as described above and even less with the 37 C/28 C regimen initiated at anthesis (0.10 and 0.16 at p < 0.056 and 0.004). Regimens initiated at 15 and 20 dpa were not related to the 37 C/ 17 C regimen initiated at anthesis. Like the 37 C/28 C regimen initiated at 10 dpa, a relatively large number of proteins responded to the high temperature regimens initiated at anthesis (Table 3). During the 37 C/28 C regimen 28 proteins increased and 29 decreased and during the 37 C/17 C regimen 24 proteins increased and 22 decreased. A comparison of the changes accompanying the high temperature regimens initiated at anthesis and 10 dpa revealed that 18 proteins increased and 17 decreased under all three regimens. The majority of the proteins that increased functioned in storage and stress defense. Storage proteins included two globulin Beg 1 isoforms (281, 620), seven globulin-2 isoforms (147, 155, 559, 851, 852, 561, 631), and a seed globulin (355). The stress defense proteins were barwin/PR-4 protein 4 (427), chitinase (351), LEA (551), and LTP (633). The majority of proteins that decreased functioned in carbohydrate metabolism and stress defense. Proteins with roles in carbohydrate metabolism were ADP-glucose PPase (130, 132), aldehyde dehydrogenase (154), and pyruvate Pi dikinase (31). Stress defense proteins were dehydroascorbate reductase (382), peroxidase (1009), purple acid phosphatase (312), and serpin (231). Like the 37 C/28 C regimen initiated at 20 dpa, relatively few proteins responded when this regimen was initiated at 15 dpa (Table 3). Of the 11 proteins that increased and 17 that decreased under the regimen, only 4 proteins increased and 6 decreased under the 37 C/28 C regimens initiated at 15 and 20 dpa. The proteins that increased have roles in stress/defense (341, chitinase; 427, barwin/ PR-4 protein 4; 633, LTP) and storage (1002, globulin Beg 1). The proteins that decreased are involved in carbohydrate metabolism (193, 6-phosphogluconate dehydrogenase; 310, glucose and ribitol dehydrogenase), storage (631 and 852, globulin-2), protein synthesis (361, 20S proteosome a-subunit), and stress/defense (312, purple acid phosphatase). 4. Discussion 4.1. Endosperm development under the moderate temperature regimen The protein accumulation proles developed in this study provide a dynamic picture of biological events that occur during wheat grain ll under a moderate temperature regimen. During
early grain ll, endosperm proteins have roles in seven functional processes: carbohydrate metabolism, nitrogen metabolism, protein synthesis/assembly, protein turnover, stress/defense, storage, signal transport, and transcription/translation. Based on the number of proteins, carbohydrate metabolism, protein synthesis/ assembly, stress/defense, and transcription/translation are the principal functions at this stage of development. During mid grain ll, the number of stress/defense and storage proteins remains at earlier levels, but the number of proteins functioning in carbohydrate metabolism, nitrogen metabolism, protein synthesis/ assembly, protein turnover, and transcription/translation decreases. Late in grain ll, the number of proteins that function in stress/defense and storage increase and proteins involved in nitrogen metabolism, signal transduction, and transcription/ translation decrease to undetectable levels. This developmental program ensures that the endosperm synthesizes the requisite reserves for germination and seedling growth as well as proteins that protect these reserves from abiotic stresses and biotic invasion during grain desiccation and dormancy. Transcript analysis using micro arrays (Laudencia-Chingcuanco et al., 2007) and serial analysis of gene expression (SAGE) (McIntosh et al., 2007) revealed a similar sequence of events during wheat grain development. A more detailed picture of grain development emerges when the levels of specic proteins within the functional categories are compared. Early in grain ll, proteins involved in carbohydrate metabolism include those that function in glycolysis (glyceraldehyde-3-P dehydrogenase, phosphoglycerate kinase, phosphoglycerate mutase, enolase, pyruvate Pi dikinase), the citric acid cycle (malate dehydrogenase, aconitase), and starch biosynthesis (ADPand UDP-glucose PPase, phosphoglucomutase). During mid grain ll, enzymes functioning in glycolysis and the citric acid cycle were proportionately less abundant. ADP glucose PPase is present at lower levels, but UDP-glucose PPase and phosphoglucomutase decrease to undetectable levels. These decreases coincide with the decline in starch accumulation that accompanies this stage of grain development. Late in grain ll, proteins involved in carbohydrate metabolism function in starch (sucrose synthase) and glucose (bamylase, glucose and ribitol dehydrogenase) degradation rather than starch biosynthesis. Isoforms of glyceraldehyde-3-P dehydrogenase are again present. Since the wheat grain is undergoing desiccation at this developmental stage, this result is consistent with an earlier report that dehydration strongly increases the level of glyceraldehyde-3-P dehydrogenase in leaves of the resurrection plant, Craterostigma plantagineum, a species capable of withstanding severe desiccation (Velasco et al., 1994). However,
W.J. Hurkman et al. / Journal of Cereal Science 49 (2009) 1223 Table 3 Proteins that increased or decreased 1.8-fold or more in grain produced under high temperature regimens initiated at anthesis, 10, 15, or 20 dpa Spot no. 420a 275a 591 110 444 869 163 347 381 871 351 282 341 551 427 1001 318 1002 1003 281 620 147 155 559 851 852 561 631 355a 633 557a 193 132 130 154a 1004 310 31 158 72 98 138a 362 33 149 225 65 118 361 205 382a 1009 312a 246 639a 231 441 285 330 301 60 69 550 236 422 418 842 Protein name Function Order no. A CM CM NM PS PS PS PT S S S S S S S S S SP SP SP SP SP SP SP SP SP SP SP SP T U CM CM CM CM CM CM CM CM NM NM NM PS PS PS PS PS PS PT PT S S S S S S SP SP ST ST TT TT CM NM PS S S 164 162 161 91 30 180 49 163 104 186 150 151 149 155 156 153 157 189 190 165 169 166 168 183 184 185 187 170 160 159 154 61 117 118 11 45 173 70 32 16 37 10 76 114 51 146 19 8 74 92 89 133 124 112 144 111 143 52 83 28 68 15 158 110 34 126 152 Cluster no. 15 15 15 5 1 28 1 15 5 33 13 13 13 13 13 13 14 36 36 15 18 16 18 31 31 32 34 19 15 14 13 2 5 5 1 1 21 2 1 1 1 1 3 5 1 11 1 1 3 5 5 7 5 5 9 5 9 1 3 1 2 1 14 5 1 5 13 37 C/28 C 10 dpa Nucleoside diphosphate kinase Glyceraldehyde 3-P dehydrogenase (NAD) Glucose and ribitol dehydrogenase 2-Isopropylmalate synthase 40S Ribosomal protein S21 60S Ribosomal protein L12 Heat shock associated protein 20S Proteasome a-subunit [proteasome subunit a type 6] a-Amylase/subtilisin inhibitor Barwin [pathogenesis-related protein 4] Chitinase-a [chitinase-c] Chitinase-a [chitinase-c] Chitinase-c Late embryogenesis abundant protein-like Pathogenesis-related protein 4 [Barwin] Unknown protein [similar to late embryogenesis abundant protein] Xylanase inhibitor protein I Globulin Beg 1 Globulin Beg 1 Globulin Beg 1 [embryo globulin] Globulin Beg 1 [embryo globulin] Globulin-2 Globulin-2 Globulin-2 Globulin-2 Globulin-2 Globulin-2 Globulin-like protein [globulin-2] Seed globulin Nonspecic lipid-transfer protein Embryo-specic protein 6-Phosphogluconate dehydrogenase ADP-glucose PPase, LS ADP-glucose PPase, LS Aldehyde dehydrogenase Aldolase Glucose and ribitol dehydrogenase Pyruvate Pi dikinase UDP-glucose PPase 10-Formyltetrahydrofolate synthetase [formate tetrahydrofolate ligase] Acetohydroxyacid synthase Ketol-reductoisomerase Elongation factor 1-b Elongation factor 2 Heat shock associated protein Heat shock protein 70 Heat shock protein 70 [DNAK-type molecular chaperone HSC70] Protein disulde isomerase 20S Proteasome a-subunit 26S Proteasome Dehydroascorbate reductase Peroxidase Purple acid phosphatase Serpin Serpin Serpin WZS2 Avenin N9 Legumin-like protein 14-3-3 Protein Guanine nucleotide-binding protein Poly(A)-binding protein Poly(A)-binding protein [formate tetrahydrofolate ligase] Sucrose synthase type 2 Aspartate amino transferase 40S Ribosomal protein S12 Superoxide dismutase [CuZn] Late embryogenesis abundant protein-like 2.09 5.15 2.79 2.57 2.58 1.81 1.83 2.74 1.87 8.29 2.39 2.21 2.54 3.00 4.22 1.92 1.86 8.33 5.41 8.46 3.83 1.85 4.08 3.35 2.92 3.86 3.78 2.07 3.23 2.53 2.41 2.20 2.63 2.47 2.26 4.92 3.34 1.99 4.36 2.52 3.25 7.30 2.85 2.79 2.01 5.20 2.08 4.83 1.93 2.09 5.66 5.34 2.27 2.98 1.92 1.85 2.41 2.50 3.97 2.80 2.88 20 dpa 2.25 37 C/28 C Anthesis 37 C/17 C Anthesis
21
37 C/28 C 15 dpa
2.33 1.83 2.52
Bb C C B , C
C C C C C C C C
C C B , C
B C C
C C C
1.85 1.82
C C C C C
1.83 C C C C C C C , B B B C B B B
2.59 2.75
2.38
6.43 2.54 3.74
B C C C C C C C C C C C C B B ,
C B B B
C C C C
C ,
B
C , B B
B B , B B
, B , , B B B
B B B
, B B
B B
, C B , B
B B
1.81
2.85 2.14 1.95
B B
B ,
, B , B
B B B
1.87 2.03 1.85 4.35 2.32 2.11
B ,
, C C
B B
C
Identities of the proteins in these spots were conrmed by MS. Accession numbers, percent coverage, expectation values, and peptide sequence data are listed in Table S2. Brackets denote current database names for proteins identied in Vensel et al. (2005). Spot 69, previoulsy identied as poly(A)-binding protein, was identied as formate tetrahydrofolate ligase in this study. XML les that contain the mass spectrometry data are available online (see Section 2.3). a When reanalyzed by MS/MS, these spots contained peptides from one or two additional proteins, see Table S2. b Symbol key: B, decrease 1.8-fold or more; C, increase 1.8-fold or more; ,, absent under high temperature regimen; -, absent under moderate temperature regimen.
22
a relationship between increased levels of this enzyme and drought tolerance has not been established. The major wheat storage proteins, the gliadins and glutenins, comprise more than 80% of endosperm protein and are found in the KCl-insoluble fraction. However, the KCl-soluble fraction contains a number of proteins with best matches to a variety of storage proteins, including avenin, triticin, legumin-like protein, seed globulin/19 kDa globulin, embryo globulin, globulin-2, or globulin Beg1. Avenin is a protein in oats that is similar to wheat gliadin. Triticin is a protein in wheat that is similar to pea legumin, an 11S globulin. Globulin-2 and globulin Beg1 are so named because they are similar to 7S globulins present in maize and barley, respectively. Like the gliadins and glutenins (DuPont et al., 2006a,b), the majority of these non-gluten storage proteins are most abundant late in grain ll. Eighteen globulins (one isoform each of avenin and globulin-like protein; two isoforms of gobulin Beg1, three of embryo globulins, six of globulin-2, ve of seed globulin) were detected during late development while only two globulins (one isoform each of avenin and legumin-like protein) were detected during early development and ve (three isoforms of avenin and one each of globulin-2 and triticin) during mid development. The complement of globulin storage proteins is very complex and warrants further analysis, particularly since some of these proteins increase in response to high temperature. A range of stress/defense proteins is present throughout endosperm development, a nding in accord with transcript prole analyses (McIntosh et al., 2007). A number of these proteins protect cells from reactive oxygen species (ROS) that are produced under normal growth conditions and can cause oxidative damage to proteins, DNA, and lipids. Thus, ascorbate peroxidase, peroxidase and GSH-dependent dehydroascorbate reductase accumulate early in grain ll. SGT1, a component of R-gene triggered disease resistance, and serpin, a serine protease inhibitor, are also present and may protect the developing grain against various pathogens. During mid grain ll, ROS-scavenging enzymes include dehydroascorbate reductase, peroxidase, catalase, and glyoxalase. Glyoxalase protects cells against reactive 2-oxoaldehydes generated by the triosephosphate isomerase reaction in glycolysis. Additional stress/ defense proteins present at this stage of development include the a-amylase inhibitors and bifunctional a-amylase/trypsin inhibitors, which guard against digestive enzymes of insects and fungi. The levels of these inhibitors are actually much higher than those reported in this study, because many of these proteins partition into the KCl-soluble/methanol-soluble fraction (Wong et al., 2004). Late in grain ll, ROS-scavenging enzymes include peroxidase and superoxide dismutase. Pathogen resistance proteins present at this stage include serpin, chitinase, which hydrolyzes the structural carbohydrate of fungal cell walls, barwin/PR-4 protein, which is induced by fungal pathogens and binds chitin, and xylanase inhibitor protein, which inhibits a fungal enzyme that degrades plant cell walls. As the developing grain matures it undergoes desiccation, an event accompanied by the accumulation of LEA proteins, proteins known to increase in response to drought stress. 4.2. Effect of high temperature on endosperm development The duration of grain ll is shortened by high temperature. Grain maturity is achieved earlier and is accompanied by corresponding decreases in kernel weight. These observations are in agreement with earlier studies in which maximum fresh weight, dry weight, and protein and starch content were lower in grain produced under high temperature regimens (37 C/17 C or 37 C/28 C) imposed from anthesis (Altenbach et al., 2002, 2003; DuPont et al., 2006b; Hurkman et al., 2003). Comparison of protein proles revealed that although the high temperature regimens compressed the developmental program, they did not substantially alter it. Analysis of protein
proles across physiologically equivalent stages under the three temperature regimens showed similar shifts in the timing of cellular functions during the transition from early to late development. The results in this study suggest that specic protein responses depend on when the high temperature regimens are imposed during development. Fewer proteins responded when the high temperature regimens were initiated during mid development (15 or 20 dpa) than during early development (anthesis or 10 dpa). Transient increases and decreases in protein levels were observed following transfer of plants to the high temperature regimens. These transient changes occurred over a period of days rather than hours as is typically observed in transcript studies. Under the regimen initiated at 10 dpa, HSP16.9, HSP70, peroxidase, triosephosphate isomerase, and globulins (legumin-like protein, seed globulin), increased transiently. The responses of HSP16.9 and HSP70 were also observed in previous studies on the effect of high temperature on wheat grain composition (Majoul et al., 2003, 2004; Skylas et al., 2002). Since ROS production is elevated during abiotic and biotic stresses (reviewed, Suzuki and Mittler, 2006), an increase in peroxidase would be expected. The increase in triosephosphate isomerase agrees with ndings that transcript (Minhas and Grover, 1999) and protein levels (Yan et al., 2005) for this enzyme increase in rice seedlings exposed to high temperature. The transient increase in globulins is interesting because it suggests that these proteins may have functional roles other than storage. One isoform of ADP-glucose PPase decreased transiently under the high temperature regimen initiated at 10 dpa, but several isoforms of the large subunit of this enzyme decreased and remained at low levels following initiation of this high temperature regimen. These results are in keeping with the decrease in transcript levels for this enzyme and the reduction of starch accumulation that accompanies high temperatures during grain ll (Altenbach et al., 2003; Hurkman et al., 2003). Under the high temperature regimen initiated at 20 dpa, the only protein that increased transiently was an isoform of HSP16.9. This nding agrees with the results of Skylas et al. (2002), where a number of HSP16.9 isoforms that increased during heat shock were absent in the mature grain. Among the proteins that responded to the high temperature regimens, seven were identied previously (Wong et al., 2003) as potential targets of thioredoxin, a widely distributed small disulde protein that functions in redox regulation. These targets function in stress/defense (a-amylase/subtilisin inhibitor 381, serpin 231, 246, 639, and dehydroascorbate reductase 382), storage (globulin Beg1 620), and carbohydrate metabolism (pyruvate Pidikinase 31). A previous study showed that a number of potential thioredoxin targets responded to drought stress, including serpin and dehydroascorbate reductase (Hajheidari et al., 2007). 4.3. High temperature and our quality A striking feature of the high temperature regimens is the increase in the levels of stress/defense and globulin storage proteins in the endosperm of the mature grain. The stress/defense proteins include those known to respond to desiccation (glyceraldehyde-3-P dehydrogenase, LEA protein) and pathogen invasion (chitinase, xylanase inhibitor protein, a-amylase/subtilisin inhibitor, barwin/ PR-4 protein, LTP, polyubiquitin, 20S proteosome a-subunit). Altenbach et al. (2007a,b) also reported that expression of transcripts for two of these genes, LTP and the PR-4 protein wheatwin, was enhanced under the 37 C/28 C regimen. The protein changes that occur late in grain development are of particular interest, because they may impact the functional properties of the resultant our. Some of the proteins that increase in response to high temperature are present in dough liquor, a soluble fraction of wheat dough, or foams prepared from dough liquor (Salt et al., 2005) and may have roles in gas bubble stabilization in dough and crumb structure of
23
bread. These include LTP, a-amylase inhibitor, xylanase inhibitor, chitinase, serpin, HSP70, DNAK-type chaperone, peroxidase, glyceraldehyde-3-P dehydrogenase, nucleoside diphosphate kinase, and globulin-2. Some of these proteins, notably a-amylase inhibitor (Posch et al., 1995; Weiss et al., 1997), glyceraldehyde-3-P dehydrogenase, triosephosphate dehydrogenase, and serpin (Sander et al., 2001), also react with sera from patients with Bakers asthma. LTP has been shown to react with sera from patients with immunoglobulin E-mediated food allergies, including wheat-dependent atopic dermatitis (Battais et al., 2005). In addition, many of the KClsoluble globulins and stress/defense proteins have structural similarities to food allergens (Jenkins et al., 2005). Further studies are warranted on the possible roles of these proteins in our quality and their allergenic potential. Acknowledgements The authors thank Drs. F.M. DuPont and R. Thilmony for critical reading of the manuscript. Appendix. Supplementary material Supplementary data associated with this article can be found in the online version, at doi:10.1016/j.jcs.2008.06.014. References
Altenbach, S.B., DuPont, F.M., Kothari, K.M., Chan, R., Johnson, E.L., Lieu, D., 2003. Temperature, water and fertilizer inuence the timing of key events during grain development in a US spring wheat. Journal of Cereal Science 37, 920. Altenbach, S.B., Kothari, K.M., 2004. Transcript proles of genes expressed in endosperm tissue are altered by high temperature during wheat grain development. Cereal Science 40, 115126. Altenbach, S.B., Kothari, K.M., Lieu, D., 2002. Environmental conditions during wheat grain development alter temporal regulation of major gluten protein genes. Cereal Chemistry 79, 279285. Altenbach, S.B., Kothari, K.M., Tanaka, C.K., Hurkman, W.J., 2007a. Expression of 9 kDa lipid transfer protein genes in developing wheat grains is enhanced by high temperature but not by post-anthesis fertilizer. Journal of Cereal Science, doi:10.1016/j.jcs.2007.03.010. Altenbach, S.B., Kothari, K.M., Tanaka, C.K., Hurkman, W.J., 2007b. Genes encoding the PR-4 protein wheatwin are developmentally regulated in wheat grains and respond to high temperature during grainll. Plant Science 173, 135143. Battais, F., Courcoux, P., Poplineau, Y., Kanny, G., Moneret-Vautrin, D.A., DeneryPapini, S., 2005. Food allergy to wheat: differences in immunoglobulin Ebinding proteins as a function of age and symptoms. Journal of Cereal Science 42, 109117. Craig, R., Cortens, J.P., Beavis, R.C., 2004. An open source system for analyzing, validating and storing protein identication data. Journal of Proteome Research 3, 12341242. DuPont, F.M., Hurkman, W.J., Vensel, W.H., Chan, R., Lopez, R., Tanaka, C.K., Altenbach, S.B., 2006a. Differential accumulation of sulfur-rich and sulfur-poor wheat our proteins is affected by temperature and mineral nutrition during grain development. Journal of Cereal Science 44, 101112. DuPont, F.M., Hurkman, W.J., Vensel, W.H., Tanaka, C.K., Kothari, K.M., Chung, O.K., Altenbach, S.B., 2006b. Protein accumulation and composition in wheat grains: effects of mineral nutrients and high temperature. European Journal of Agronomy 25, 96107. Hajheidari, M., Elivazi, A., Buchanan, B.B., Wong, J.H., Majidi, I., Salekdeh, G.H., 2007. Proteomics uncovers a role for redox in drought tolerance in wheat. Journal of Proteome Research 6, 14511460.
Hurkman, W.J., McCue, K.F., Altenbach, S.B., Korn, A., Tanaka, C.K., Kothari, K.M., Johnson, E.L., Bechtel, D.B., Wilson, J.D., Anderson, O.D., DuPont, F.M., 2003. Effect of temperature on expression of genes encoding enzymes for starch biosynthesis in developing wheat endosperm. Plant Science 164, 873881. Hurkman, W.J., Tanaka, C.K., 2004. A simplied method for the isolation of wheat endosperm proteins for proteome analysis. Journal of Cereal Science 40, 294299. Jenkins, J.A., Grifths-Jones, S., Shewry, P.R., Breiteneder, H., et al., 2005. Structural relatedness of plant food allergens with specic reference to cross-reactive allergens: an in silico analysis. Journal of Allergy and Clinical Immunology 115, 163170. Jung, H.W., Kim, W., Hwang, B.K., 2003. Three pathogen-inducible genes encoding lipid transfer protein from pepper are differentially activated by pathogens, abiotic, and environmental stresses. Plant, Cell, and Environment 26, 915928. Laudencia-Chingcuanco, D.L., Stamova, B.S., You, F.M., Lazo, G.R., Beckles, D.M., Anderson, O.D., 2007. Transcriptional proling of wheat caryopsis development using cDNA microarrays. Plant Moleular Biology 63, 651668. Majoul, T., Bancel, E., Triboi, E., Ben Hamida, J., Branlard, G., 2003. Proteomic analysis of the effect of heat stress on hexaploid wheat grain: characterization of heatresponsive proteins from total endosperm. Proteomics 3, 175183. Majoul, T., Bancel, E., Triboi, E., Ben Hamida, J., Branlard, G., 2004. Proteomic analysis of heat stress on hexaploid wheat grain: characterization of heat-responsive proteins from non-prolamins fraction. Proteomics 3, 505513. McIntosh, S., Watson, L., Bundock, P., Crawford, A., White, J., Cordeiro, G., Barbary, D., Rooke, L., Henry, R., 2007. SAGE of the developing wheat caryopsis. Plant Biotechnology Journal 5, 6983. Minhas, D., Grover, A., 1999. Transcript levels of genes encoding various glycolytic and fermentation enzymes change in response to abiotic stresses. Plant Science 146, 4151. Posch, A., Weill, W., Wheeler, C., Dunn, M.J., Gorg, A., 1995. Sequence analysis of wheat grain allergens separated by two-dimensional electrophoresis with immobilized pH gradients. Electrophoresis 16, 11151119. Salt, L.J., Robertson, J.A., Jenkins, J.A., Mulholland, J., Clare Mills, E.N., 2005. The identication of foam-forming soluble proteins from wheat (Triticum aestivum) dough. Proteomics 5, 16121623. Sander, I., Flagge, A., Merget, R., Halder, T.M., Meyer, H.E., Baur, X., 2001. Identication of wheat our allergens by means of 2-dimensional immunoblotting. Journal of Allergy and Clinical Immunology 107, 907913. SAS Institute Inc., 2004. SAS OnlineDoc 9.1.3. SAS Institute Inc., Cary, NC. Skylas, D.J., Cordwell, S.J., Hains, P.G., Larsen, M.R., Basseal, D.J., Walsh, B.J., Blumenthal, C., Rathmell, W., Copeland, L., Wrigley, C.W., 2002. Heat shock of wheat during grain lling: proteins associated with heat-intolerance. Journal of Cereal Science 35, 175188. Suzuki, N., Mittler, R., 2006. Reactive oxygen species and temperature stresses: a delicate balance between signaling and destruction. Physiologia Plantarum 126, 4551. Velasco, R., Salamini, F., Bartels, D., 1994. Dehydration and ABA increase mRNA levels and enzyme activity of cytosolic GAPDH in the resurrection plant Craterostigma plantagineum. Plant Molecular Biology 26, 541546. Vensel, W.H., Tanaka, C.K., Cai, N., Wong, J.H., Buchanan, B.B., Hurkman, W.J., 2005. Developmental changes in the metabolic protein proles of wheat endosperm. Proteomics 5, 15941611. Weiss, W., Huber, G., Engel, K.-H., Pethran, A., Dunn, M.J., Gooley, A.A., Gorg, A., 1997. Identication and characterization of wheat grain albumin/globulin allergens. Electrophoresis 18, 826833. Wong, J.H., Balmer, Y., Cai, N., Tanaka, C.K., Vensel, W.H., Hurkman, W.J., Buchanan, B.B., 2003. Unraveling thioredoxin-linked metabolic processes of cereal starchy endosperm using proteomics. FEBS Letters 547, 151156. Wong, J.H., Cai, N., Tanaka, C.K., Vensel, W.H., Hurkman, W.J., Buchanan, B.B., 2004. Thioredoxin reduction alters the solubility of proteins of wheat starch endosperm: an early event in cereal germination. Plant Cell Physiology 45, 407415. Wu, G., Robertson, A.J., Liu, X., Zheng, P., Wilen, R.W., Nesbitt, N.T., Gusta, L.V., 2004. A lipid transfer protein gene BG-14 is differentially regulated by abiotic stress. Journal of Plant Physiology 161, 449458. Yan, S., Tang, Z., Su, W., Sun, W., 2005. Proteomic analysis of salt stress-responsive proteins in rice root. Proteomics 5, 235244. Zadoks, J.C., Chang, T.T., Konzak, C.F., 1974. A decimal code for the growth stages of cereals. Weed Research 14, 415421.

Accumulation of mixed linkage (1 / 3) (1 / 4)-b-D-glucan during grain lling in barley: A vibrational spectroscopy study
Helene Fast Seefeldt a, b, Andreas Blennow c, Birthe Mller Jespersen b, Bernd Wollenweber a, Sren Balling Engelsen b, *
University of Aarhus, Faculty of Agricultural Sciences, Department of Genetics and Biotechnology, Forsgsvej 1, DK-4200 Slagelse, Denmark University of Copenhagen, Faculty of Life Sciences, Department of Food Science, Quality & Technology, DK-1958 Frederiksberg C, Denmark University of Copenhagen, Faculty of Life Sciences, Department of Plant Biology and Biotechnology, VKR research centre Pro-Active Plants, DK-1871 Frederiksberg C, Denmark
b c a
Article history: Received 20 November 2007 Received in revised form 9 June 2008 Accepted 30 June 2008 Keywords: Infrared IR Near-infrared NIR (1 / 3) (1 / 4)-b-D-glucan Grain lling
a b s t r a c t
The accumulation of mixed linkage barley (1 / 3) (1 / 4)-b-D-glucan (BG) during grain lling at eight stages was studied using standard reference methods and infrared spectroscopy. Two mutant barley genotypes having higher (starch mutant lys5f) and lower (high lysine mutant lys3a) BG content than the normal control Cork were studied. The Cork and lys3a genotypes showed a linear BG accumulation throughout the grain lling to reach a maximum of approximately 6 and 4% BG (w/w) dry matter, respectively. However, lys5f mutant exhibited an exponential increase in BG synthesis to a maximum of approximately 18% BG (w/w) dry matter 30 days after owering (DAF), seemingly compensating for a decreased synthesis of starch. The spectral information of the barley our was compared to pure BG spectra and partial least squares regression (PLS) models were constructed for calibration to BG content. Informative regions in the nearinfrared (NIR) and the infrared (IR) spectra were identied for separation of temporal and genetic differences. Interval PLS yielded good calibration models to BG (R2 0.94 for NIR in the region 11941240 nm, whereas the global PLS gave correlations with BG with R2 0.92 for IR). 2008 Elsevier Ltd. All rights reserved.
1. Introduction In the quest for optimising cereals for soluble bres and other health-promoting components, there is great interest in studying the metabolic changes during grain lling, such as cell wall bre development in barley mutants differing in mixed linkage (1 / 3) (1 / 4)-b-D-glucan (BG) using high throughput methods. Inexpensive, spectroscopic methods based on vibrational spectroscopy offer the possibility of fast and exible analysis of a large number of genotypes (Osborne, 2006). Near-infrared (NIR) and infrared (IR) spectroscopy measure the vibrations of molecular covalent bonds. The NIR region 14,3004000 cm1 (7802500 nm) mainly gives
Abbreviations: ADP-glucose, adenosine 50 diphosphate glucose; ATR, attenuated total reection; BG, Mixed linkage (1 / 3) (1 / 4)-b-D-glucan; DAF, days after owering; EISC, extended inverted signal correction; FT-IR, Fourier transform infrared; iPLS, interval partial least squares regression; MSC, multiplicative scatter correction; NIR, near-infrared reectance; PC, principal component; PCA, principal component analysis; PLS, partial least squares regression; RMSECV, root mean square error of cross validation; SECV, standard error of cross validation; SEE, standard error of estimate; VIS, visual part of the NIR spectra. * Corresponding author. Tel.: 45 3533 3205; fax: 45 3533 3245. E-mail address: se@life.ku.dk (S.B. Engelsen). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.06.012
information about overtones and combination tones (stretching and bending) involving anharmonic bonds primarily to hydrogen, whereas the IR region 4000200 cm1 (250050,000 nm) gives information about the fundamental vibrations. NIR spectroscopy is already well established for at/on-line quality control in the food and food ingredients industries (Zachariassen et al., 2005) and is able to provide information about chemical parameters such as water, protein and starch content as well as about physical parameters such as particle size and temperature. In cereals, NIR transmittance spectra of single seeds have proven very informative for the determination of different quality traits such as protein and fat content with high accuracy (Delwiche 1995; Pedersen et al., 2002). In fact, a new high capacity single-seed TriQ NIR sorting system (Bomill AB, Lund, Sweden) (Munck, 2007) has been developed utilising NIR spectra to diversify heterogeneous bulk lots of wheat with regard to multivariate complex quality traits such as dough performance and baking value (Munck, 2007; Tnning, 2008). Furthermore, NIR can be used to evaluate quality traits such as the bre fraction of cereal cell walls (Blakeney and Flinn, 2005) and genetics (Jacobsen et al., 2005; Munck et al., 2004). While NIR spectroscopy has primarily been used in quantitative analysis of bulk components (in spite of its documented ability to predict complex qualitative traits like baking and malting quality
H.F. Seefeldt et al. / Journal of Cereal Science 49 (2009) 2431
25
(Munck, 2007)), IR spectroscopy has mainly been used to study well-dened components such as plant cell wall polysaccharides (Chen et al., 1998; Kacurakova and Wilson, 2001; McCann et al., 1992; Robert et al., 2005; Sene et al., 1994). Recently, micro Fourier transformed infrared spectroscopy (FT-IR) was used to study the deposition of cell wall polysaccharides in wheat endosperm during grain development with emphasis on BG and arabinoxylans (Philippe et al., 2006). Also, NIR spectroscopy has proven valuable in monitoring plant physiological processes such as carbohydrate accumulation during grain lling (Gergely and Salgo, 2005). NIR screenings of normal barley and mutants have resulted in the discovery of genotypes with a strongly increased content of soluble BG bres (Munck et al., 2004). The high BG barley mutant line lys5f is a structural (enzyme functional) low starch mutant (Munck et al., 2004; Munck and Mller, 2005) unable to transport ADPglucose across the plastid envelope due to an inactive ADPglucose transporter (Patron et al., 2004). Lys5f is thus disabled in efcient synthesis of starch, but compensates via an extremely high content (approximately 17% d.m.) of soluble and insoluble BG (Munck et al., 2004). The lys3a mutant is a regulatory mutant (Jacobsen et al., 2005) that primarily inhibits the synthesis of the hordein, but increases the synthesis of soluble proteins. It has approximately 2 3% BG compared to 34% in normal barley like Cork when grown in green houses (Munck et al., 2004). The aim of this study was thus to characterise and investigate two extreme recessive barley mutants with respect to BG and starch content during grain lling using classical reference methods as well as spectroscopic ngerprinting methods. The carbohydrate mutant lys5f with high BG content and the protein lys3a with low BG content were compared to the variety Cork that has a normal content of BG. 2. Experimental 2.1. Plant material Three genotypes of barley were included in the study: a malt barley (Hordeum vulgare cv. Cork) and a barley mutant lys3a with alterations in the lys3 locus on chromosome ve, tightly linked to adjacent BG synthesis suppressing genes (Munck et al., 2004). The mutant lys5f has a mutation in chromosome six. A semield pot experiment (72 pots, 16.5 cm diameter, 13 cm height) was carried out from April to August 2005 at the University of Aarhus, Research Centre Flakkebjerg, Denmark. Each pot was lled with 10 l of sphagnum with 15% Perlite (Perlite, Denmark) added. Ten seeds of each genotype were sown on April 20th and thinned to three seedlings per pot on May 31st. All pots were drip-watered throughout the experiment and standard pest control was performed against mildew and aphids when needed. Flowering was judged visually when 50% of the spikes showed clear pollen release, which occurred between the 26th and 27th of June. Spikes were harvested at eight time points during grain lling: 9, 13, 16, 20, 23, 30, 39 and 47 days after owering (DAF). 2.2. Plant analysis The spike on the main tiller and the rst spike of the side tillers were cut and immediately frozen in liquid nitrogen. After freezing, the kernels were detached from the spike, counted and weighed. Kernels were transferred to 80 C and freeze-dried within 3 weeks after harvest. One sample consisted of the seeds from two spikes from the same individual plant. The freeze-dried grains were milled (0.5 mm, Cyclotec 1093, Foss Tecator AB, Hogenas, Sweden). A total of 91 samples were analysed (Cork, 31; lys5f, 31; and lys3a, 29). The total sample set consisted of three genotypes at eight temporal harvest points two replicate spectra of each genotype from
harvests 14 and six replicate spectra of each genotype from harvests 58 were analysed. This experimental design yielded a total of 96 samples (3 genotypes 4 (14 harvests) 2 replicates 3 genotypes 4 (58 harvests) 6 replicates). Due to a minor aw in the experiment, three replicates of lys3a from harvests 1, 5 and 7, and two replicates of Cork harvest 5 and lys5f harvest 5 were lost. The ground material was stored in sealed plastic bottles at room temperature until analysis. 2.3. Chemical analysis The content of soluble BG was analysed by uorimetry (Calcoour reagent type II, Scandinavian Brewery Laboratory, Frederiksberg, Denmark) (Munck et al., 1989). The high BG content ours showed major deviations in values when determined by Calcoour, and hence the values from the last four time points in lys5f were double-checked with an enzymatic kit (Megazyme, Wicklow, Ireland) specic for mixed linkage BG. The total content of starch was analysed as follows: freeze-dried our (10 mg) was washed three times with 1-ml aliquots of 80% ethanol in screw-cap Eppendorf tubes. Starch in the washed our was gelatinised by the addition of KOH (400 ml, 0.2 M) and incubation at 95 C for 1 h. After cooling, 140 ml of 1 M acetic acid was added, the samples mixed and diluted 20-fold with water. The diluted sample (10 ml) was mixed with an equal volume of amyloglucosidase (10 U/ml, Fluka) solubilised in 50 mM Na acetate pH 5.0 and starch was hydrolysed by incubation at 37 C for 2 h. A 210-ml aliquot of 50 mM Mops/KOH pH 7.3, 5 mM MgCl2, 1 mM EDTA, 1 mM ATP, 1 mM NAD and 4 U/ml hexokinase was added and absorbance at 340 nm was registered. One microlitre of 500 U/ml glucose-6-phosphate dehydrogenase was added and the production of NADH was followed until steady from the absorbance at 340 nm. Starch content in the original our was calculated using glucose as standard. Amylose in the our was determined by iodine complexation, as described by Bay-Smidt et al. (1999). The amylopectin chain length distribution of the starch was analysed by high performance anionic exchange chromatography with pulsed amperometric detection (HPAEC-PAD) as described by Blennow et al. (1998). FT-IR and NIR spectra were obtained from the pure substances: Cellulose (9004d34-6, Sigma-Aldrich Chemie, Steinheim, Germany), BG (Barley Medium Viscosity, Megazyme, Wicklow, Ireland) and wheat starch with normal and high content of amylose (Ritmo, Sejet Plantbreeding, Horsens, Denmark). The starches were puried according to Blennow et al. (1998). 2.4. FT-IR measurements All FT-IR spectra were acquired at room temperature using the Arid-Zone MB100 FT-IR spectrometer (ABB Bomen Inc., Quebec, PQ, Canada). The sampling was performed using an Attenuated Total Reection (ATR) device with a diamond crystal (ZnSe, TR-plate, ARK 0055-603, Spectra-Tech Inc., CT, USA) operating in the range 4000 750 cm1. Measurements were obtained using 64 scans at 4 cm1 resolution. Background scans were obtained using 128 scans. The scans were averaged and ratioed against a single-beam spectrum of the clean ATR crystal and converted into absorbance units. 2.5. NIR measurements A near-infrared instrument (Foss NIRSystems 6500, USA) was used in reectance mode in the range of 4002500 nm. All measurements were acquired at room temperature using a commercial small sample cup containing approximately 0.5 g of our. The spectra were recorded in 2-nm steps using a spinning sample module. Each sample was measured using 32 scans which
26
were ratioed against 16 reference scans using a ceramic sample. The results were averaged as a log 1/R spectrum. 2.6. Data analysis The spectra were analysed using the chemometric software LatentiX 1.0 (http://www.latentix.com, Latent5, Copenhagen, Denmark) both for visual inspection of the spectra and calculation of principal component analysis (PCA) (Wold et al., 1987) and partial least squares (PLS) regressions. The PCA was performed to visualise systematic spectral variation. Calibrations and predictions of BG and amylose based on the spectral information were made using PLS. Interval partial least squares regression (iPLS) (Nrgaard et al., 2000) was performed in order to reveal the most important spectral interval correlated with the calibration parameters. Prior to analysis, the spectral data was scatter-corrected using the extended inverted signal correction (EISC) (Martens et al., 2003; Pedersen et al., 2002) and all models are mean-centred. The iPLS was carried out using the iTOOLBOX (http://www.models.life.ku.dk) as a routine for MATLAB 6.0 (The Mathworks Inc., Natick, USA). All reported models are based on 30 intervals and are validated using full cross validation. 3. Results and discussion
reaching a maximum of approximately 47% starch remaining at this level until 39 DAF. It should be noted that the levels of grain starch at maturation in our study were somewhat lower than generally found in barley grain as a result of a decrease in starch content of approximately 10% the last 17 DAF. The mutant lys5f displayed a much slower increase throughout the entire grain lling period until 39 DAF to reach a maximum at 23% starch. A low nal level of starch in lys5f was previously reported (Munck et al., 2004). The high content of BG in lys5f can hence be interpreted as a new mechanism for compensating starch synthesis with BG synthesis (Munck et al., 2004). Interestingly, the total accumulation of starch and BG reveals three similar patterns for the genotypes with just an offset in total accumulation. Thus, the rapid accumulation rate of BG in lys5f corresponds largely to the rapid accumulation of starch in the non-starch genotypes lys3a and Cork. This conrms the observation of the starch and BG levels found in mature seed described by (Munck et al., 2004). The higher starch content found in Cork was followed by a high dry matter weight (Fig. 1e). The lys5f mutant had a markedly lower ratio of amylose/amylopectin (Fig. 1f) which is interesting from a starch functionality viewpoint (Tester et al., 2004). The chain length of amylopectin did, however, not differ between the three genotypes and it did not differ markedly over the developmental stages (data not shown). 3.2. Near-infrared analysis of the barley grain lling
The seeds showed normal fresh weight and water content during development (Fig. 1a,b) during grain lling (Gergely and Salgo, 2003; Jennings and Morton, 1962): Three distinct stages of the grain lling process could be determined by analysis of the fresh grain weight (Fig. 1a). Until 20 DAF a rapid increase in grain weight occurred, followed by a lag phase with stable grain weight. After 30 DAF the grain weight decreased due to the drying and maturation of the seed. The rst phase is characterised by a rapid inux of water and cell enlargement and continues until 1015 DAF. During the second phase the seeds take up nutrients and start to synthesise starch and protein (Jenner et al., 1991). In the third phase the fresh weight decreases due to loss of water as indicated in Fig. 1b. 3.1. Accumulation of BG during grain lling chemical analysis Fig. 1c shows the development of BG synthesis between genotypes during grain lling. The analysis of the BG content of the three barley genotypes (Fig. 1c) revealed that lys5f from 16 to 30 DAF exhibited a very strong increase in BG accumulation to reach an extreme maximum BG content of approximately 18% at 30 DAF. During the same period lys3a and Cork exhibited slow linear accumulation of BG, which continued throughout the grain lling period until maturity at 47 DAF. The mutant lys3a had lower BG content as compared to both Cork and lys5f, reaching only approximately 4% at harvest. Cork displayed a normal content of BG, reaching a maximum level of approximately 6% BG at maturity. During grain lling, Cork and lys3a did not start to accumulate BG until 1620 DAF, whereas lys5f already started to accumulate BG between 9 and 13 DAF (Fig. 1c). In a previous study of three normal barley varieties (Coles, 1979), a large increase in accumulation of BG starting at 19 DAF was found. In wheat, the cell walls mainly consist of arabinoxylans, whereas BG constitutes only 25%, but BG was accumulating from early grain lling until 10 DAF after which arabinoxylans were dominating (Philippe et al., 2006). These observations led to the hypothesis that BG could function as a structural element of growing cell walls as well as storage material that could be hydrolysed during germination of the grain (Philippe et al., 2006). The starch synthesis in Cork and lys3a (Fig. 1d) followed the same pattern having a rapid increase from 9 DAF until 20 DAF and
The EISC pre-processed VIS/NIR spectra (4002498 nm) of the 91 barley our samples are displayed in Fig. 2a. The rst three harvests (916 DAF) represented by the light blue colours clearly differ in their NIR spectral proles from the rest of the harvests (20 47 DAF, shown as dark red colours). The rst three harvests are less intensive in the rst overtone region from 14001850 nm and the O-H combination tone centred at 1940 nm. In general, the temporal differences inuenced the offset in the spectra, while the genetic differences inuenced the shape of the spectra. In the visible part of the spectrum, the grain lling is clearly evidenced by the gradual elimination of the chlorophyll peak at 672 nm (Fig. 2b), which reects that the barley grains were visibly green until 30 DAF (the sixth harvest). Fig. 2c displays a score plot from a PCA of the 91 NIR spectra of the barley our, explaining 92% of the variance. In the score plot, the temporal variation is separated into four welldened groups. The rst three harvests (916 DAF) are markedly different from the later harvests and located in the upper left corner of the PCA. The fourth and the fth harvests (20 and 23 DAF), in which the BG is also primarily synthesised (Fig. 1c), are clustered with the lowest scores in the score plot, the sixth harvest (30 DAF) located in the lower right part of the spectra and the last two harvests (3947 DAF) in the upper right corner of the PCA. By inspection of the loadings, the primary reason for the clear separation of the temporal grain llings was found to be the chlorophyll peak at 672 nm (Table 1). As observed in Fig. 2c, the actual time of owering is difcult to determine in the gure a few samples reside in their previous group (marked with stars in Fig. 2c), presumably due to inaccuracy in the DAF determination. Munck et al. (2004) found that the combination tone region 22802360 nm contains unique information about the lys3 mutants and Szczodrak et al. (1992) found that two wavelengths in the region between 2260 and 2380 nm correlate the best with BG. Fig. 2d shows that the NIR spectra of lys5f have a peak in the NIR spectra around 2345 nm (the second bold arrow). This peak is also present in the spectra of lys3a, but as a shoulder, whereas the spectra of Cork only have a very weak shoulder. On the other hand, the spectra of Cork have a characteristic shoulder at 2280 nm, indicating starch (rst bold arrow in Fig. 3d, Table 1), according to Munck (2007). This shoulder is absent in the mutants lys5f and lys3a. As Cork and lys3a have almost identical starch contents, it
27
110.0 95.0
80.0
mg seed-1
60.0
65.0 50.0 35.0 9 13 16 20 23 30 39 47
%
40.0 20.0 9 13 16 20 23 30 39 47
80.0
Fresh weight
22.0 20.0 18.0 16.0 14.0 45.0
Water content
60.0
% dm
12.0 10.0 8.0 6.0 4.0 2.0 0.0 9 13 16 20 23 30 39 47
% dm
30.0
15.0
0.0 9 13 16 20 23 30 39 47
Beta-glucan
60.0 50.0 40.0
Starch
40.0 35.0 30.0
mg seed-1
% starch
9 13 16 20 23 30 39 47
30.0 20.0 10.0 0.0
25.0 20.0 15.0 10.0 9 13 16 20 23 30 39 47
Dry matter
Amylose
Fig. 1. (A) Fresh seed weight (mg) (B) Water content (%). (C) BG content (% d.m.). (D) Starch content (% d.m.). (E) Seed dry weight (mg). (F) Amylose (% in starch). All values are denoted in relation to DAF (days after owering). Solid, Cork; dotted, lys5f; broken, lys3a.
could be questioned whether this peak corresponds to starch only. When comparing the average barley our NIR spectra with the NIR spectra of pure cellulose, starch, BG and amylose (Fig. 3a), it is evident that a high degree of similarity exists with the starch spectrum except for the region 22002400 nm where the spectra show very distinct proles including genotype specic patterns. From the lines indicated in Fig. 2d it can be observed that the three genotypes exhibit similar spectral properties during the rst harvest (9 DAF) as well as during the second and third harvests. In order to avoid the dominance of the chlorophyll information, the visual part of the spectra was discarded and the NIR region 1100 2498 nm was examined separately (data not shown). The three genotypes are clustered together until 16 DAF. From the fourth harvest (20 DAF) and onwards, the three genotypes are separated into three distinct clusters. This is in excellent agreement with the
rapid increase in the BG synthesis from 16 DAF to 23 DAF for lys5f (Fig. 1c) and the concomitant rapid increase in starch seen for lys3a and Cork. The loadings associated with the PCA reveal that the rst PC is mainly spanned by information from moisture and starch (OH stretching vibrations, 1440 and 1940 nm) (Table 1), which is in good accordance with the separation of grain lling time points along PC1. The third PC separates the genotypes into a Cork/lys3a cluster and a lys5f cluster. The corresponding loadings are starchrelated peaks (Munck et al., 2004; Szczodrak et al., 1992), which manifest the different syntheses in the three genotypes. 3.3. Infrared investigation of barley grain lling The EISC-treated FT-IR spectra (1900750 cm1) of the 91 barley our samples are shown in Fig. 4a. As was the case for the NIR
28
0.55 0.5 0.45
0.36 0.34 0.32
Log 1/R
0.4 0.35 0.3
Log 1/R
800 1200 1600 2000 2400
0.30 0.28 0.26
0.25 0.24 0.2 400 580 620 660 700 740
nm
1 2 3 4 5 6 7 8
nm
Levels of Harvest
0.15 0.10
1 1 2 2 1 2 2 1 2 23 3 3 3 3 3 78 7 88 7 7 8 88 8 7 8 877 887 58 77 88 77778 8 6 6 6 6 6 7 6 66 6 6 6 6 6 6 66 6 5 5 5 5 55 4 4 4 5 55 5 6 4 5 45 5
Scores PC#2 (14.%)
0.05 0.00 -0.05 -0.10 -0.15 -0.20 -0.5
1log/R
9 DAF Cork lys5f lys3a 0.3

2290 2310 2330 2350
-0.4
-0.3
-0.2
-0.1
0.1
0.2
Scores PC#1 (78.%)
nm
Fig. 2. The complete NIR spectra (4002498 nm) of the three genotypes and eight harvest times. (A) The full, EISC-treated spectra. Squares indicate enlargements shown in B and D. (B) Magnication of the chlorophyll peak (672 nm). (C) PCA model of the full NIR region. Blue, rst three harvests; Purple, fourth and fth harvests; cerise, sixth harvest; pink, seventh and eighth harvests. (D) The region 22802360 nm. Open arrows indicate spectral features of interest. Black arrow indicates spectra of 9 DAF.
spectra, a clear temporal dependence is observed in the spectra (data not shown), but the difference between genotypes appears more pronounced than in the NIR spectra. The barley our spectra represent typical broad absorption bands of polysaccharides in the region 1200750 cm1 (Kacurakova and Wilson, 2001) with a maximum absorption around 1021 cm1. The glycosidic absorption band at 1160 cm1 reported by Robert et al. (2005) and at 1150 cm1 (Philippe et al., 2006) is found at 1152 cm1 in our samples. Fig. 4c (raw spectra) shows the enlargement of the maximum absorption region in which a clear
difference between the three genotypes is observed. At 1002 cm1 the high BG mutant lys5f has a different intensity compared to Cork and lys3a. Moreover, lys5f has a lower intensity of the shoulder around 1070 cm1, whereas lys3a and Cork have well-dened peaks at 1078 cm1. With regard to the temporal changes, a shift from 1036 cm1 for 9 DAF towards 1021 cm1 for the end of the grain lling is observed in the spectra (Fig. 4c). The anomer specic peaks in the range from 800 to 950 cm1 contain information about BG as well as other polysaccharide cell wall components (Philippe et al., 2006). Fig. 4d displays an enlargement of this region as
Table 1 The most informative peaks and their putative alignments as determined from loading plots corresponding to PCAs made on the dened regions Region NIR method 500750 500750 11002498 11002498 22802360 22802360 IR method 1900750 1900750 PC 1 2 1 3 1 2 1 2 Peak 1 672 672 1444 2250 2308 2288 1010 1049 Peak 2 400 400 1856 2477 2348 2314 960 1147 Peak 3 1440 500 1930 2355 2360 Peak 4 1930 560 1670 2384 Alignment 672: chlorophyll a, 1440: CH, OH in starch,1940 H2O 1440 OH starch, H2O, 1940 H2O 2252, 2461 OH starch, 2353 CH cellulose, 2380: ROH 2310 CH2, 2347 HCCHCH2 2280: CH3, 2310 CH CH2, 2352: CH cellulose 900990: aromatic, 10001260: CO alcohol, 1002 ring stretching (arabinoxylansa), 11601210:CC(O)C ester
1002
1094
The peaks with the highest intensity are listed rst. indicates a positive loading, while indicates a negative loading. Units for region: NIR, nm; IR, cm1. a From (Philippe et al., 2006). All other alignments from (Osborne et al., 1993).
29
A
0.45 Average NIR flour spectra Betaglucan Cellulose 0.4 Wheat starch Wheat amylose 0.35 0.12 0.14 0.16
B
Average IR flour spectra Betaglucan Cellulose Wheat starch Wheat amylose
Absorbance
logR/1
0.3
0.1 BG peak
0.25
0.08
0.2
0.06
0.15
0.04
0.1
0.02
1400
1600
1800
2000
2200
2400
800 900 1000 1100 1200 1300 1400 1500 1600 1700
nm
cm-1
Fig. 3. IR and NIR spectra of pure seed carbohydrates compared to an averaged IR and NIR our spectrum. (A) IR spectra. The arrow indicates the BG and cellulose specic peak. (B) NIR spectra.
0.11 0.1 0.09 0.08
Cork
lys5f lys3a
0.10 0.08
B
7 8 7 67 8 8 2 7 656 8 6 6 55 3 2 5 3 47 7 6 4 2 68 2 2 8 8 7 78 3 6 6 7 55 3 6 878 6 4 75746 3 3 76 6 778 8 5 55 4 88 4 7 87 5 65 6 6 5 5
1 1
Scores PC#2 (1.4 %)
0.06 0.04 0.02 0 -0.02 -0.04 -0.06
Absorbance
0.07 0.06 0.05 0.04 0.03 0.02 0.01 800 1000 1200 1400 1600 1800
Cork + lys5f lys3a
-0.36 -0.34 -0.32 -0.3 -0.28 0.26 -0.24 -0.22 -0.2
cm-1
0.11 0.10
Scores PC#1 (97 %)
Cork lys5f lys3a
Absorbance
0.08 0.07 0.06 0.05
9-13 DAF
2.derivative
0.09
Cork lys5f lys3a 980 1000 1020 1040 1060 1080 1100 1120 820 840 860 880 900 920 940
cm-1
cm-1
Fig. 4. FT-IR spectra of the region 7501900 cm1. (A) EISC-treated spectra. Squares are indicated magnied in C and D. (B) PCA of the second derivative of the spectra (not meancentred). (C) EISC-pre-processed spectra of the region of maximum absorption. Arrows indicate 9 and 13 DAF. (D) The second-derivative spectra of the region from 800 to 940 cm1.
30
6 5
B
20
r = 0.97 RMSECV = 1.57
Predicted (BG%)
3 1 4 5 6 8 7 7 5 6 1 7 6 4 5 6 5 4 6 2 561 5 4 86 5 3 1
RMSECV
4 3 2 1
15 10 5 0
1098
1290
1482
1672
1856
2040
2224
2408
10
15
20
Wavelength(nm)
Measured (BG%)
Fig. 5. RMSECV correlation coefcients of iPLS models of NIR and FT-IR spectra. (A) The iPLS performed on 30 intervals superimposed on the NIR spectra. Many intervals are better calibrators (smaller RMSECV values) for BG than the global model using two PLS components. Numbers on bars indicate PLS components used. The red marked interval 1194 1240 nm is the best calibrator. (B) Measured versus predicted plot of the model in the interval 11941240 nm and with three outliers expelled: two replicates of lys5f, 30 DAF and one lys5f at 9 DAF.
second-derivative spectra which reveal that lys5f has less intensive peaks compared to Cork and lys3a. Cork and lys3a display a relatively strong a-anomer band at 855 cm1, whereas the high BG lys5f and the early stages of grain lling show a much weaker a-glucan peak. This is in good accordance with the starch levels in Cork and lys3a compared to lys5f. In the case of the b-anomer sensitive peak normally present at approximately 890 cm1 (Fig. 4d), we observed higher background intensity in lys5f than in the other two genotypes. From approximately 23 DAF, the intensity is the highest corresponding to the highest level of BG. Pure barley BG shows a peak at 895 cm1, not present in the average grain spectra (see Fig. 3b). A PCA was applied to the full second-derivative IR spectra (Fig. 4b) and the rst PC explains 97% of the variance. Both the rst and the second PC separate the genotypes. The corresponding loading plot (Table 1) revealed that PC1 is mainly spanned by the peak at 1010 cm1 and PC2 at 1049 cm1, which are the dominating peaks of starch in the second-derivative spectra of the pure substances (Fig. 3b). Thus BG seems to play a less pivotal role in discriminating the genotypes compared to starch.
revealed that the region 11941240 nm yielded an R2 of 0.94 and RMSECV of 1.6% using four PLS components (Fig. 5a,b). This subregion covers second overtones of the C-H stretching and shows that BG can be predicted well in this region. PLS calibrations to the FT-IR region and BG were also studied. The global FT-IR models predicted BG slightly better than the global NIR model with a squared correlation coefcient of 0.92 and a RMSECV value of 1.83 using six PLS components and removing two outliers (lys5f from 9 DAF and from 30 DAF) from the model. Subsequently an iPLS was applied to the IR spectra but no intervals were able to improve the calibration performance when compared to the global model. It thus appear that the information regarding BG is hidden across the IR spectra from 1900 to 750 cm1, and not conned to the beta specic peak at approx 890 cm1, whereas the BG information in the NIR spectra can with advantage be correlated with the region of C-H stretching at 11941240 nm. Correlations with the amylose content were predicted with a squared correlation coefcient of 0.81 and an RMSECV of 2.46 for the FT-IR and 0.77 and an RMSECV of 2.71 for NIR. The poorer correlations with amylose probably reect the uncertainty of the amylose determinations more than the ability of the spectra to predict amylose or starch content.
3.4. Relationships between the chemical and spectroscopic methods In order to study the relationships between the spectral data and grain lling, a range of calibration models were investigated. Global PLS models based on the entire data set were used, as well as iPLS models, where sub-regions are tested against the global model. Only a few studies have attempted to construct an NIR calibration to BG content. The quality of the predictions has varied greatly. A correlation coefcient of 0.69 and an RMSE of 0.557% was achieved in the BG range 5.88.4% (Czuchajowska et al., 1992). In another study, an 1-VR value that is a determination for cross validation and measure of goodness of t was found to be 0.92 and an SECV of 0.45 in the BG range 0.095.12% (Blakeney and Flinn, 2005). One major problem has been the limited BG range for the calibrations. However, Szczodrak et al. (1992) succeeded in developing calibration models to BG with a correlation coefcient of 0.871 and an SEE (standard error of estimate) of 0.677 using the calibration range 2.79.5%. In this study, the BG variation (calibration range) was further expanded in order to establish a quantitative model between the NIR spectra and the BG content. The global NIR model predicts BG with a squared correlation coefcient of 0.84 and an RMSECV of 2.60 using two PLS components with three outliers removed (lys5f from 9 DAF and two from 30 DAF). However, application of iPLS 4. Conclusions A rapid increase in grain fresh weight during grain lling is concomitant with the increase of starch and BG. The high BG mutant lys5f shows an exponential increase of BG until 23 DAF and reaches a nal content of 18%. This is in contrast to the low BG mutant lys3a and the conventional malting barley Cork having linear phases of BG deposition and reaching maxima of only 4 and 6%, respectively. Starch deposition in these latter genotypes was exponential, reaching maxima of 47% compared to a linear increase in lys5f leading to a maximum of 23% starch. The NIR and FT-IR spectra are complete chemical ngerprints of the grain and thus more complex to resolve than single chemical analysis of grains. Data extraction relies on the use of chemometrics to reveal hidden information. In the case of NIR, the genetic variation was mainly resolved from the spectral region between 2260 and 2380 nm. Moreover, the NIR spectra showed mostly temporal differences primarily due to complex changes in moisture. In the NIR region 22602380 nm, all the barley our spectra differed and the average grain spectrum was inuenced by both the starch/ amylose and the BG contents, which underlines the importance of this NIR region for breeding purposes.
31
Both NIR and IR spectra gave good correlations with BG. The NIR sub-region 11941240 nm gave the best correlation to BG with an R2 of 0.94 which is promising for using NIR to predict BG in heterogeneous samples such as seeds and our with a wide span in BG content. Fast and non-destructive spectroscopic ngerprinting is obviously advantageous for studying physiological processes such as grain lling, as the techniques are non-destructive, fast and sensitive and it is possible to do real-time analysis. Acknowledgements The DIAS Competence Fund and the Ministry of Food Agriculture and Fisheries are greatly acknowledged for nancial support to the project (FFS05-9: Build Your Food). Betina Srensen (University of Aarhus) and Lisbeth T. Hansen and Louise Nancke (Faculty of Life Sciences, University of Copenhagen) are acknowledged for technical support in semi-eld experiments and chemical analysis. Gilda Kischinovsky is acknowledged for proofreading. The authors wish to thank Lars Munck for valuable comments and for sharing his unique barley mutant collection. References
Bay-Smidt, A.M., Blennow, A., Bojko, M., Mller, B.L., 1999. The effect of phosphate and amylopectin molecular size on peak viscosity of starch pastes. Annual Transactions of the Nordic Rheology Society 7, 3138. Blakeney, A.B., Flinn, P.C., 2005. Determination of non-starch polysaccharides in cereal grains with near-infrared reectance spectroscopy. Molecular Nutrition & Food Research 49, 546550. Blennow, A., Bay-Smidt, A.M., Wischmann, B., Olsen, C.E., Mller, B.L., 1998. The degree of starch phosphorylation is related to the chain length distribution of the neutral and the phosphorylated chains of amylopectin. Carbohydrate Research 307, 4554. Chen, L., Carpita, N.C., Reiter, W.-D., Wilson, R.H., Jeffries, C., McCann, M.C., 1998. A rapid method to screen for cell-wall mutants using discriminant analysis of Fourier transform infrared spectra. The Plant Journal 16, 385392. Coles, G., 1979. Relationship of mixed-link beta-glucan accumulation to accumulation of free sugars and other glucans in the developing barley endosperm. Carlsberg Research Communications 44, 439453. Czuchajowska, Z., Szczodrak, J., Pomeranz, Y., 1992. Characterization and estimation of barley polysaccharides by near-infrared spectroscopy.1. Barleys, starches, and betaDglucans. Cereal Chemistry 69, 413418. Delwiche, S.R., 1995. Single wheat kernel analysis by near-infrared transmittance: protein content. Cereal Chemistry 72, 1116. Gergely, S., Salgo, A., 2003. Changes in moisture content during wheat maturation what is measured by near infrared spectroscopy? Journal of Near Infrared Spectroscopy 11, 1726. Gergely, S., Salgo, A., 2005. Changes in carbohydrate content during wheat maturation what is measured by near infrared spectroscopy? Journal of Near Infrared Spectroscopy 13, 917. Jacobsen, S., Sndergaard, I., Mller, B., Desler, T., Munck, L., 2005. A chemometric evaluation of the underlying physical and chemical patterns that support near infrared spectroscopy of barley seeds as a tool for explorative classication of endoperm genes and gene combinations. Journal of Cereal Science 42, 281299. Jenner, C.F., Ugalde, T.D., Aspinall, D., 1991. The physiology of starch and protein deposition in the endosperm of wheat. Australian Journal of Plant Physiology 18, 211226.
Jennings, A.C., Morton, R.K., 1962. Changes in carbohydrate, protein and non-protein nitrogenous compounds of developing wheat grain. Australian Journal of Biological Sciences 16, 318331. Kacurakova, M., Wilson, R.H., 2001. Developments in mid-infrared FT-IR spectroscopy of selected carbohydrates. Carbohydrate Polymers 44, 291303. Martens, H., Nielsen, J.P., Engelsen, S.B., 2003. Light scattering and light absorbance separated by extended multiplicative signal correction. Application to nearinfrared transmission analysis of powder mixtures. Analytical Chemistry 75, 394404. McCann, M.C., Hammouri, M., Wilson, R., Belton, P., Roberts, K., 1992. Fourier transform infrared microspectroscopy is a new way to look at plant cell walls. Plant Physiology 100, 19401947. Munck, L., 2007. Breeding for Quality Traits in Cereals A Revised Outlook on Old and New Tools for Integrated Breeding. Cereals. Springer Verlag, New York. Munck, L., Mller, B., 2005. Principal component analysis of near infrared spectra as a tool of endosperm mutant characterisation and in barley breeding for quality. Czech Journal of Genetics and Plant Breeding 41, 8995. Munck, L., Jrgensen, K.G., Ruudhansen, J., Hansen, K.T., 1989. The EBC methods for determination of high molecular-weight beta-glucan in barley, malt, wort and beer. Journal of the Institute of Brewing 95, 7982. Munck, L., Mller, B., Jacobsen, S., Sndergaard, I., 2004. Near infrared spectra indicate specic mutant endosperm genes and reveal a new mechanism for substituting starch with (1-3,1-4) betaglucan in barley. Journal of Cereal Science 40, 213222. Nrgaard, L., Saudland, A., Wagner, J., Nielsen, J.P., Munck, L., Engelsen, S.B., 2000. Interval partial least-squares regression (iPLS). A comparative chemometric study with an example from near-infrared spectroscopy. Applied Spectroscopy 54, 413419. Osborne, B.G., 2006. Applications of near infrared spectroscopy in quality screening of early-generation material in cereal breeding programmes. Journal of Infrared Spectroscopy 14, 93101. Osborne, B.G., Fearn, T., Hindle, P.H., 1993. Theory of near infrared spectrophotometry. Practical NIR Spectroscopy with Applications in Food and Beverage Analysis, second ed. Longman Scientic and Technical, Harlow, UK, pp. 13-35. Patron, N.J., Greber, B., Fahy, B.F., Laurie, D.A., Parker, M.L., Denyer, K., 2004. The lys5 mutations of barley reveal the nature and importance of plastidial ADP-Glc transporters for starch synthesis in cereal endosperm. Plant Physiology 135, 20882097. Pedersen, D.K., Martens, H., Nielsen, J.P., Engelsen, S.B., 2002. Near-infrared absorption and scattering separated by extended inverted signal correction (EISC): analysis of near-infrared transmittance spectra of single wheat seeds. Applied Spectroscopy 56, 12061214. Philippe, S., Robert, P., Barron, C., Saulnier, L., Guillon, F., 2006. Deposition of cell wall polysaccharides in wheat endosperm during grain development: Fourier transform-infrared microspectroscopy study. Journal of Agricultural and Food Chemistry 54, 23032308. Robert, P., Marquis, M., Barron, C., Guillon, F., Saulnier, L., 2005. FT-IR investigation of cell wall polysaccharides from cereal grains. Arabinoxylan infrared assignment. Journal of Agricultural and Food Chemistry 53, 70147018. Sene, C.F.B., McCann, M.C., Wilson, R.H., Grinter, R., 1994. Fourier-transform raman and Fourier-transform infrared spectroscopy an investigation of ve higher plant cell walls and their components. Plant Physiology 106, 16231631. Szczodrak, J., Czuchajowska, Z., Pomeranz, Y., 1992. Characterization and estimation of barley polysaccharides by near-infrared spectroscopy. 2. Estimation of total betaDglucans. Cereal Chemistry 69, 419423. Tester, R.F., Karkalas, J., Qi, X., 2004. Starch composition, ne structure and architecture. Journal of Cereal Science 39, 151165. Tnning, E., 2008. Wheat baking quality in a process analytical perspective. Sampling, diversication, prediction and chemometric method development. PhD thesis, University of Copenhagen. ISBN: 87-91949-23-8. Wold, S., Esbensen, K.H., Geladi, P., 1987. Principal component analysis. Chemometrics and Intelligent Laboratory Systems 2, 3752. Zachariassen, C.B., Larsen, J., van den Berg, F., Engelsen, S.B., 2005. Use of NIR spectroscopy and chemometrics for on-line process monitoring of ammonia in low methoxylated amidated pectin production. Chemometrics and Intelligent Laboratory Systems 76, 149161.

Mechanism of gas cell stabilization in bread making. I. The primary glutenstarch matrixq
Baninder S. Sroan a, *, Scott R. Bean b, Finlay MacRitchie a
a b
Department of Grain Science and Industry, Kansas State University, Manhattan, KS 66506, USA USDA-ARS, Grain Marketing and Production Research Center, Manhattan, KS 66502, USA
Article history: Received 19 December 2007 Received in revised form 28 June 2008 Accepted 8 July 2008 Keywords: Bread making Gas cell stability Liquid lamella Glutenstarch matrix Dough rheology
a b s t r a c t
Expansion of dough and hence bread making performance is postulated to depend on a dual mechanism for stabilization of inating gas bubbles. Two ours were used in this study, one from the wheat variety Jagger (Jagger) and the other from a composite of soft wheat varieties (Soft). Thin liquid lamellae (lms), stabilized by adsorbed surface active compounds, act as an auxiliary to the primary glutenstarch matrix in stabilizing expanding gas cells and this mechanism operates when discontinuities begin to appear in the glutenstarch matrix during later proving and early baking stages. Contributions of the liquid lamellae stability to dough expansion were assessed using ours varying in their lipid content. Incremental addition of natural lipids back into defatted our caused bread volume to decrease, and, after reaching a minimum, to increase. Strain hardening is a key rheological property responsible for stabilizing the primary glutenstarch matrix. Jagger gave higher test-bake loaf volume than Soft and higher strain hardening index for dough. The different lipid treatments were found to have negligible effects on strain hardening index. Image analysis of crumb grain revealed that differences in number of gas cells and average cell elongation with different lipid treatments were insignicant. The evidence agrees with a dual mechanism to stabilize the gas cells in bread dough. To understand dough rheology at a molecular level, rheological properties of doughs were varied by addition of our protein fractions prepared by pH fractionation. Fractions were characterized by SE-HPLC and MALLS. The molecular weight distribution (MWD) of fractions progressively shifted to higher values as the pH of fractionation decreased. Mixograph dough development time paralleled the MWD. However, the strain hardening index and the testbake loaf volume increased with increasing MWD up to a point (optimum), after which they declined. At a given strain rate, the behavior at the optimum is thought to result from slippage of the maximum number of statistical segments between entanglements, without disrupting the entangled network of polymeric proteins. Shift of MWD to molecular weight higher than the optimum results in a stronger network with reduced slippage through entanglement nodes, whereas a shift to lower molecular weights will decrease the strength of the network due to a lesser number of entanglements per chain. 2008 Elsevier Ltd. All rights reserved.
1. Introduction During mixing, gas is occluded and concentrated in the liquid phase of dough in the form of small nuclei. No further occlusion of
Abbreviations: MALLS, multiangle laser light scattering; MDDT, mixograph dough development time; MT, threshold molecular weight; MW, molecular weight; MW, weight average molecular weight; MWD, molecular weight distribution; Rmax, extensograph maximum resistance to extension; SDS, sodium dodecyl sulfate; SEHPLC, size-exclusion high-performance liquid chromatography; UPP, unextractable polymeric protein. q Mention of rm names or trade products does not constitute endorsement by the U.S. Department of Agriculture over others not mentioned. * Corresponding author. Present address: FritoLay, 7701 Legacy Dr., 3T-124, Plano, TX 75024, USA. Tel.: 1 972 334 4526; fax: 1 972 334 2329. E-mail address: baninder@k-state.edu (B.S. Sroan). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.07.003
gas occurs in succeeding stages (sheeting, molding, etc.) of the bread making process (Baker and Mize, 1946). However, these subsequent stages (punching, sheeting and molding) cause subdivision of already existing gas cells, thus improving their number and size distribution. Gas nuclei expand during fermentation due to release of fermentation gases, and during baking due to expansion of these gases as temperature increases. Wheat our dough, due to its unique visco-elastic properties, is capable of stabilizing the expanding gas cells. The classical view has emphasized only the glutenstarch matrix as the cell membrane that stabilizes the expanding gas cells (Bloksma, 1990; Hoseney, 1992). However, scanning electron micrographs of dough at the end of the proving stage appear to show the existence of intact gas cells even when discontinuities in the glutenstarch matrix begin to appear (Gan et al., 1990). Based on these electron micrographs of
B.S. Sroan et al. / Journal of Cereal Science 49 (2009) 3240
33
the dough, Gan and co-workers (Gan et al., 1990, 1995) proposed that the expanding gas cells appear to be stabilized against coalescence and disproportionation by a primary glutenstarch matrix with a secondary liquid lamella on its inner side, enveloping the gas cell. The dual lm hypothesis seems plausible in view of evidence from previous studies, which show presence of a continuous liquid phase in dough (MacRitchie, 1976a) and the effect of surface active components on bread loaf volume (DeStefanis and Ponte, 1976; MacRitchie, 1976a,b; MacRitchie and Gras, 1973). In such cases, where there is an independent mechanism of gas cell stabilization apart from the glutenstarch matrix, the dough rheology should not be affected by variations in natural lipid levels. MacRitchie and Gras (1973) reported that Alveograms are not affected by variations in the natural lipid fraction of the our. A Stable Micro Systems dough ination system mounted on a texture analyzer (TAXT 2plus), used in this study to understand biaxial extensional rheology of wheat our dough, works on a principle similar to the Alveograph, with the advantage of inating bubbles at constant strain rates. The amounts of lipid in the study by MacRitchie and Gras (1973) were as low as 11.5% and had signicant effects on baking performance. It is probably due to their surface activity that these compounds (lipids and proteins) are adsorbed at the gas liquid interface of the liquid lamellae and affect stability of the gas cells (Mills et al., 2003; Paternotte et al., 1993; Ross and MacRitchie, 1995). Analogous variations in foaming properties of the dough liquor, and bread loaf volume and crumb structure (MacRitchie, 1976a) provide reasonable evidence for their action at the interface. Loaf volume can be dened as the extent of dough expansion (Gandikota and MacRitchie, 2005), which depends upon how thin the glutenstarch matrix can be stretched before reaching its expansion limit or point of rupture. Rheology of the glutenstarch matrix is important in the bread making process as this determines extensibility and strength. The proving and baking stages of bread making are characterized by fast biaxial expansion of gas cells, expanding at strain rates of 0.0010.0001/s and 0.010.001/s, respectively (Dobraszczyk, 1997). During expansion, the gluten starch matrix around gas cells expands biaxially to large strains (>100%) due to excess pressure produced in the gas cells by diffusion of carbon dioxide during proving, and by thermal expansion of gases during baking. This causes thinning of gas cell walls and, if a gas cell continues to expand along this thin region, it may rupture. However, if the stress in the thin region increases more than proportionally to strain, the thin region of a cell wall or (glutenstarch matrix) will resist further deformation and the gas cell will continue to expand along thicker parts of the cell wall. This localized increase of stress in response to strain, preventing failure of the gas cell walls, is called strain hardening (Dobraszczyk and Roberts, 1994; van Vliet et al., 1992), and is a necessary rheological property for obtaining good bread volume. A power law relation between stress and Hencky strain, as described by the equation below, was used by Dobraszczyk and Roberts (1994) to determine strain hardening.
Fig. 1. Stress vs. strain (Hencky) curves of Soft and Jagger wheat ours showing differences in their tendencies to strain harden. The Soft wheat our with poor bread making potential fails at relatively lower strains.
s [ K 3n
where, s is the true stress, K is a power law constant, 3 is strain and n is a strain hardening index. The index n must be greater than 1 in order to have a curved relation between stress and Hencky strain (Fig. 1). Doughs with a strain hardening index of 1 or higher have the potential to give good loaf volume by allowing gas cells to expand without undergoing disproportionation and coalescence (Dobraszczyk and Roberts, 1994). In later studies, Dobraszczyk and co-workers (Dobraszczyk and Salmanowicz, 2008; Dobraszczyk et al., 2003) found that an exponential relationship shows better t to data, particularly in the case of doughs for which bubbles inate
to strains greater than 2. Chin and Campbell (2005) also used an exponential relationship to study dough rheology by bubble ination. Strain hardening index is a reliable criterion for differentiating ours based on bread making potential (Dobraszczyk and Roberts, 1994; Dobraszczyk and Salmanowicz, 2008; Dobraszczyk et al., 2003; Tronsmo et al., 2003;). This is illustrated in Fig. 1 for the two ours used in this study. At a molecular level, strain hardening can be explained by a well established polymer entanglement network theory (MacRitchie and Laandra, 1997; Singh and MacRitchie, 2001). Strain hardening is believed to originate from entanglement coupling of large glutenin molecules (Singh and MacRitchie, 2001), with molecular weight (MW) greater than a threshold MW (MT), MT being the minimum molecular weight at which stable entanglements are formed (Bersted and Anderson, 1990). Various studies substantiate this theory. Dough strength in terms of Rmax (extensograph maximum resistance to extension) shows positive correlation of the glutenin fraction with MW greater than MT, which is estimated to be approximately 250,000 Da (Bangur et al., 1997). The shift in MW distribution (MWD) of this fraction towards higher MW or increase in relative proportion of this fraction will increase the number of entanglements per chain and reduce the MW between entanglements (Termonia and Smith, 1988). This will lead to an increase in strength (Gupta et al., 1990). A similar phenomenon is observed in wheat our doughs, where the strain hardening index (n) seems to have a positive curvilinear relation with failure strain. As n approaches higher limits, the failure strain seems to become constant (Dobraszczyk and Roberts, 1994; Dobraszczyk et al., 2003). Successive addition of glutenin fractions with increasing molecular weight to a base our at constant protein level causes loaf volume to increase, and, after reaching a maximum, to decrease (Lundh and MacRitchie, 1989; MacRitchie, 1987; MacRitchie et al., 1991). The optimum here possibly indicates a balance between strength and extensibility, beyond which high entanglement network density, resulting in decrease in failure strain, might cause lower loaf volumes (Singh and MacRitchie, 2001; Termonia and Smith, 1988). The objectives of this study were: rstly, to investigate and seek evidence for the presence of liquid lamellae and their ability to stabilize gas cells; i.e. whether or not the dual lm hypothesis is plausible; secondly, to understand the rheology of the gluten starch matrix required to stabilize the expanding gas cells and to understand the underlying molecular structurefunction relationship as explained by polymer entanglement network theory (MacRitchie and Laandra, 1997; Singh and MacRitchie, 2001).
34
2. Experimental 2.1. Materials Jagger our and soft wheat our (a blend of soft wheat varieties) were two untreated and unbleached ours used in this study. The ours were milled in a Buhler mill (73% milling extraction rate) and were evaluated for certain quality characteristics which are shown in Table 1. These ours were stored at 20 C until use. For the sake of brevity, Jagger wheat our and soft wheat our have been referred to as Jagger and Soft, respectively. Reagents were purchased from SigmaAldrich, USA. Chloroform was of HPLC grade, whereas all other chemicals used were of ACS grade. Distilled deionized water, sterilized in an autoclave, was used in all stages of the experiments. 2.2. Analytical procedures Moisture content was determined as per AACC method 44-15 A (AACC, 2001). Protein content was determined by the nitrogen combustion method using a LECO FP-2000 nitrogen/protein analyzer with a factor of 5.7 to convert N to protein. Lipid content was determined by cold extraction with chloroform (HPLC grade) using the procedure of MacRitchie and Gras (1973). 2.3. Dough mixing properties Mixing properties were evaluated using a 10 g Mixograph (National Manufacturing Co., Lincoln, NE). Mixing parameters (peak development time and weakening angle) were used for comparison. The procedure used was similar to AACC method 55-40 (AACC, 2001) except that sodium chloride (1.5% w/w on a our weight basis) was added. 2.4. Preparation of gluten proteins Gluten protein was prepared from defatted Jagger our following the procedure used by MacRitchie (1985). The wet gluten mass was freeze-dried and ground to a particle size less than 150 microns. Powdered gluten was stored at a temperature of 20 C until its use. 2.5. pH fractionation of gluten protein Powdered gluten prepared from Jagger was subjected to pH fractionation, using the procedure of Gupta et al. (1990). Gluten was stirred in water at pH 5.3 (35 g/1000 ml) for 10 min and centrifuged at 2300 g for 10 min. The residue, after collection of supernatant, was subjected to two more extractions at the same pH. The total
combined supernatant was named as pH 5.3 gluten fraction. The residue was subjected to further fractionation in a similar manner by lowering of pH to 4.9 by addition of HCl. The gluten protein was thus fractionated into 6 fractions by sequential lowering of pH. Fractions were obtained at pH 5.3, 4.9, 4.1, 3.5 and 3.1, obtaining two fractions at pH 3.1 i.e. supernatant of pH 3.1 and residue of pH 3.1. All fractions were then freeze-dried. Freeze-dried fractions were ground to particle sizes less than 150 microns. The fractions were stored in polyethylene bags in containers with desiccant at a temperature of 20 C until their use. Fractions were analyzed for percentage protein and moisture. Protein size characterization of the fractions was done using SE-HPLC and MALLS. Fractions retained their functional properties (dough mixing properties) when reconstituted in proportions relative to their yield, with starch (Midsol-50, MGP Ingredients, Inc., Hutchison, KS) to form a our with protein content equal to that of the original our (Jagger). 2.6. Addition of protein fractions to our Addition of protein fractions to defatted base our (Jagger and Soft) was done at a level of 1% (dry protein) on a total our weight basis. 2.7. Size-exclusion high-performance liquid chromatography (SE-HPLC) and multiangle laser light scattering (MALLS) Size characterization of gluten proteins was done using SE-HPLC (Hewlett-Packard 1100 system) in conjunction with MALLS (Multiangle light scattering detector DAWN EOS of Wyatt Technology Corp., Santa Barbara, CA). The basic procedure for SE-HPLC is described elsewhere (Batey et al., 1991). Proteins were fractionated with a Biosep SEC-4000 column (Phenomenex, Torrance, CA) using; 50 mM disodium orthophosphate buffer (NaPhos), pH 7.0, containing 1% sodium dodecyl sulfate (SDS), as mobile phase at a ow rate of 0.5 ml/min. Proteins were detected at 214 nm. Injection volume of samples for total and extractable protein analysis was 80 ml and that for unextractable protein analysis was 40 ml. Therefore, results obtained from chromatograms of unextractable proteins were multiplied by a factor of 2. Reduced injection volume for unextractable protein analysis was used to avoid issues in MALLS analysis due to the high concentration of very large glutenin protein. SEC-HPLC data was analyzed using software program ChemStation (Agilent Technologies, USA). MALLS data was analyzed with software program ASTRA 4.50 (Wyatt Technology Corp.) using 0.31 as the dn/dc value, as per procedure by Bean and Lookhart (2001). 2.8. Sample preparation for SE-HPLC The basic procedure for SE-HPLC is based on the method of (Batey et al., 1991. Samples were prepared to analyze total, extractable and unextractable polymeric protein (Gupta et al., 1993). Samples for total and extractable protein analysis were weighed in microfuge tubes. Sample size was determined based on protein content of each sample such that protein content in all samples was kept constant, using Jagger our as standard and the quantity for Jagger being 10.0 mg 0.1 mg. Weighed samples were suspended in 50 mM NaPhos, pH 6.9, 0.5% SDS and solubilized by vortexing for 10 min. For total protein analysis, to achieve solubilization of the largest molecular size fraction, samples were sonicated (Singh et al., 1990) at room temperature at an output of 6 W for 15 s. No sonication step was included for extractable protein analysis. The sonicator probe was placed at 1/3 distance from the bottom of the microfuge tube. Microfuge tubes with protein suspensions were then centrifuged at 12,000 g for 20 min. The
Table 1 Physico-chemical analysis and SE-HPLC relative composition of polymeric proteins of Jagger and Soft wheat ours Jagger wheat our Protein (%) (14% moisture basis) Lipid Content (%) (14% moisture basis) Mixograph midline peak development time (min) Mixograph absorption (%) Area (%) under chromatogram curve TPP EPP UPP 10.4 0.02 0.89 0.03 7.21 0.06 63 36.4 0.01 43.6 0.02 56.4 2.77 Soft wheat our 9.2 0.03 0.93 0.01 9.25 0.00 60 37.0 0.05 39.0 0.11 61.1 3.40
TPPdtotal polymeric protein; EPPdextractable polymeric protein; UPPdunextractable polymeric protein.
35
supernatant was decanted in HPLC vials and sealed. To ensure stability of prepared samples, the vials with supernatant were heat treated in a water bath at 85 C for 5 min to inhibit any intrinsic proteolytic activity. After heat treatment, vials were cooled with crushed ice and analyzed by SE-HPLC. Residue from extractable protein was used for unextractable protein analysis. The same procedure was followed, except that suspensions were sonicated for 25 s at an output of 6 W. 2.9. Lipid extraction from our Flour lipids were extracted using three batch extractions with chloroform in a glass beaker, followed by Buchner ltration through Whatman No. 1 lter paper (MacRitchie and Gras, 1973). 200 g of our and 400 ml of chloroform were used for each extraction. The defatted our was spread out on a at glass tray in a fume hood for 12 h to allow evaporation of solvent. 2.10. Addition of lipids to defatted our For incorporation of intact natural our lipids into defatted our, the parent our was mixed with defatted our in different proportions, for both the ours, to give different our lipid levels. 2.11. Test baking Test loaves (35 g our) were baked using a modied rapid bake test (MacRitchie and Gras, 1973). A lean formulation was used with no added shortening; our (100%), sugar (6%), sodium chloride (1.5%), instant yeast (2.7%), potassium bromate (30 ppm), water and mix time (as optimized from Mixograph analysis). Loaf volumes were measured by rapeseed displacement after cooling for 20 min. 2.12. Image analysis Image analysis of crumb grain of baked loaves was carried out 12 h after baking, with a C-Cell, an image analyzing software and equipment (Calibre Control International Ltd., UK). Loaves were sliced using a rotary disc blade (unserrated Graef blade) cutter. Image analysis was performed on central slices of 15 mm thickness, as soon as possible to avoid any shrinkage of crumb grain. Image analysis parameters (number of cells and average cell elongation) were used for comparison between different treatments. 2.13. Biaxial extensional rheology Biaxial extensional rheological properties of the doughs were measured with a Stable Micro Systems dough ination system mounted on a texture analyzer (TAXT 2plus) by means of the procedure established by Dobraszczyk (1997). Doughs for rheological testing were mixed in the same mixer as used for bake tests, using the same water absorption, mixing times and sodium chloride addition. After mixing, dough pieces were squashed by hand on a sheeting board without putting too much stress on the dough, and then allowed to relax for 5 min. They were then sheeted, rolled out slowly with several passes and rotated by 90 degrees after each pass. Sheeting was done for 5 min with relaxation of 10 s between each pass. Sheeting in all directions prevents anisotropic effects during dough ination, allowing dough pieces to expand uniformly into spherical shapes. After sheeting, dough pieces were relaxed for 20 min. They were then cut into circular discs using a 55 mm cookie cutter, squashed to a height of 2.67 mm for 20 s. Sample dough pieces (in pots) were then proved at 35 C for 25 min. During sample preparation, dough pieces were protected against loss of moisture using a ne coating of mineral oil (Saybold viscosity 335/ 358) and covering with shrink wrap lm. Mineral oil of lower
viscosity seems to penetrate dough pieces and may affect rheological measurements. Dough pieces were inated at a ow rate of 500 cm3/min at a strain rate of 0.1/s. Rheological parameters (peak stress, failure strain and strain hardening index) were used to compare between different treatments. Strain hardening index was calculated by tting an exponential curve to the stressstrain (Hencky) curve, after transferring data to Microsoft Excel. 2.14. Statistical analysis Results were analyzed using analysis of variance (ANOVA). ANOVA was performed using a general linear model procedure to determine signicant differences and interactions for the various treatments. Means were compared by using Fishers LSD procedure (a 0.05). Statistical analysis was performed using proc GLM in SAS (version 9.1; SAS Institute Inc., Cary, NC) software. Duplicates were prepared for each treatment and the order of treatment was not signicant. 3. Results and discussion This section is further divided in two sub-sections in accordance with objectives of the study. 3.1. General mechanism of gas cell stability (investigating dual lm hypothesis) 3.1.1. Physico-chemical analysis of ours Two wheat ours, Jagger and Soft, used for this study were analyzed for their chemical composition and physical dough mixing properties. Moisture (wet weight basis), protein (14% moisture basis) and lipid (14% moisture basis) contents of the ours are given in Table 1. Jagger (10.4%) was nearly 1% higher in protein content than Soft (9.2%). However, Soft gave higher mixing time (9.25 min) compared to Jagger (7.21 min). Higher percentage of unextractable polymeric protein (UPP) (Table 1) in Soft explains the higher mixing requirements for this our. This is not typical of soft wheat ours, which normally are considered weak. However, some soft varieties like Caldwell are relatively stronger than others. Soft wheat our was milled from a composite of soft wheat varieties grown in Missouri during 20042005 and obtained from grain elevators at Kansas City, MO. Increase in mixing requirements in terms of Mixograph dough development time (MDDT) with increase in percent UPP has been reported previously by Gupta et al. (1993). Data on polymeric proteins was obtained by SE-HPLC analysis of the ours. Lipid content of Jagger and Soft was 0.89% and 0.93%, respectively. 3.1.2. Effect of natural our lipid level variation on baking performance 3.1.2.1. Bread making. Signicant differences (P < 0.0001) in loaf volumes of the two ours were observed in response to varying levels of our lipids (Fig. 2). Incremental addition of our lipids back into defatted parent our caused bread volume to decrease, and after reaching a minimum, to increase. At all levels of lipid addition, relatively lower volumes were observed for Soft compared to Jagger. For Soft, minimum volume is reached at a relatively lower percentage of natural our lipids (w30%) in comparison to Jagger (w50%) (Fig. 2). The results of loaf volume variations in response to natural our lipid level variations agreed qualitatively with those of MacRitchie and co-workers (MacRitchie, 1976a; MacRitchie and Gras, 1973; McCormack et al., 1991). Loaf volume seems to be governed by expansion capacity of the gas cells. The expansion capacity of the gas cells can be dened as the extent to which gas cells can expand
36
B.S. Sroan et al. / Journal of Cereal Science 49 (2009) 3240 Table 2 Crumb structure responses of Jagger and Soft wheat our breads to natural our lipid levels % of Jagger wheat our natural Number of cells our lipids 0 20 40 50 60 80 100 2000.0 11.3 a 1935.0 222.03 a,b 1880.5 78.5 a,b 1898.0 59.4 a,b 1788.5 81.3 a,b 1907.5 2.1 a,b 1749.0 41.0 b Soft wheat our Average cell elongation (C-cell score) 1.58 0.01 b 1.59 0.05 b 1.64 0.06 a,b 1.68 0.04 a 1.67 0.05 a,b 1.62 0.01 a,b 1.63 0.01 a,b Number of cells Average cell elongation (C-cell score) 1.69 0.01 a,b 1.68 0.01 a,b,c 1.65 0.03 b,c 1.65 0.02 c 1.70 0.02 a 1.66 0.00 a,b,c 1.69 0.03 a,b
2544.5 187.4 a,b 2521.0 110.3 a,b,c 2210.5 115.3 d 2209.0 79.2 d 2347.5 2.1 b,c,d 2196.0 86.3 d 2229.0 114.6 d
Values represent mean standard deviation for duplicate determinations. Means with the same letter within columns are not signicantly different (p > 0.05).
Fig. 2. Loaf volume vs. intact natural our lipids, added to defatted Jagger (- - -) and Soft (d) wheat ours as percentage of natural our lipids.
without failure; i.e. the value reached after which the loaf does not increase in volume. This maximum for doughs giving lower loaf volumes is achieved during the late proving stage, and for those giving higher loaf volumes is achieved during the early baking stage. This was determined by measuring oven spring (data not shown). Oven spring was observed only in doughs giving higher loaf volume. Scanning electron micrographs by Gan et al. (1990) appeared to show the presence of intact gas cells with discontinuities in gluten starch matrix at advanced stages of proving. This means that the expansion capacity of the gas cells is not just controlled by the glutenstarch matrix and there is possibly a secondary factor contributing to it. This secondary factor as hypothesized by Gan et al. (1990) is a liquid lamella present on the inner side of the gas cell. When liquid lamellae also fail, the presence of discontinuities in the glutenstarch matrix leads to coalescence of gas cells, thus decreasing volume. It is quite possible that lipids, due to their surface action, will be affecting stability of liquid lamellae, thus causing variation in loaf volume. However, this needs to be investigated further in order to make sure that the action of our lipids is independent of the rheological properties of the glutenstarch matrix. Relatively lower loaf volumes at all levels of lipid addition in Soft, could be attributed to inherent differences in gluten quality of the two ours, as previously suggested by MacRitchie (1978). The differences in the two ours are clear from SEC-MALLS data (Table 5). A new desirable relative proportion of polymeric proteins greater than 250,000 Da in Jagger is probably responsible for better loaf volumes as explained in Section 3.2.2.3. 3.1.2.2. Crumb structure. Image analysis of bread crumb (Table 2) showed that addition of different levels of our lipids resulted in insignicant variations in number of gas cells (P 0.02) and average cell elongation (P 0.94 for Jagger and P 0.12 for Soft). These non-signicant differences in number of gas cells further make it evident that differences in loaf volumes are due to differences in expansion capacity of the gas cells (Table 2), and factors like gas cell concentration are not involved. Negligible variations in average cell elongation indicated that lipids might not be causing any variations in rheology of the glutenstarch matrix, since cell elongation is thought to be associated with dough rheology (Gandikota and MacRitchie, 2005). 3.1.3. Effect of natural our lipid level variation on biaxial extensional rheology of glutenstarch matrix In order to investigate possible independent action of lipids in causing loaf volume variations, to indicate presence of liquid
lamellae, biaxial extensional rheology tests were conducted on defatted Jagger and Soft doughs, and at our lipid levels where maximum and minimum loaf volumes were observed. Biaxial extensional rheological parameters (maximum stress, failure strain and strain hardening index) were higher for Jagger doughs (Table 3) in comparison to Soft doughs. Different lipid treatments did not inuence these parameters within doughs prepared from each our type (Table 3). Though minor differences were observed for strain hardening indices and failure strain of the doughs at some lipid levels, no specic trend was observed, thus attributing the variation to slight experimental scatter. Higher values of biaxial extensional rheological parameters for Jagger verify that, rheologically, Jagger is a better bread making our. Higher strain hardening index and failure strains ensure that gas cells in Jagger doughs expand to a greater extent than Soft doughs. Similar results that differentiate between bread making potential of ours have been reported by Dobraszczyk and coworkers (Dobraszczyk and Roberts, 1994; Dobraszczyk and Salmanowicz, 2008; Dobraszczyk et al., 2003;). On the other hand, variations in natural lipid levels in a particular our that produce signicant variations in loaf volume (Fig. 2) made no difference to biaxial extensional rheology of the dough from that our. This provides clear evidence of the role of lipids as surface active compounds stabilizing liquid lamellae, a phenomenon independent of dough rheology. This study, for the rst time provides clear evidence on the presence of liquid lamellae as an independent stabilizing mechanism working auxiliary to the glutenstarch matrix in stabilizing expanding gas cells during bread making. The study also demonstrates that lipids at their natural levels do not affect biaxial extensional rheology as determined by dough ination.
Table 3 Mean bubble ination rheological response of Jagger and Soft wheat our doughs to natural our lipid levels % of natural our lipid Jagger wheat our 0 60 100 Soft wheat our 0 40 100 Max. stress (Kpa) Failure strain (Hencky) 2.58 0.18 a 2.60 0.12 a 2.66 0.04 a 1.79 0.00 a 1.85 0.02 a 1.66 0.05 b Strain hardening index 2.21 0.12 b,c 2.38 0.064 a,b 2.44 0.06 a 1.76 0.01 a 1.74 0.01 a 1.64 0.04 c
571.62 305.45 a 491.68 237.21 a 598.62 112.05 a 120.80 12.76 a 125.42 12.14 a 102.40 5.45 a
37
3.2. Mechanism of stability of the glutenstarch matrix 3.2.1. Gluten fractionation Yield (%) and protein content (%) of each fraction from the pH fractionation are given in Table 4. Cumulative percentage of total polymeric protein extracted is plotted as a function of pH of solution in Fig. 3. This was calculated using yield data and percentage total polymeric protein of each fraction as given in Table 4. Approximately 35% of total polymeric proteins and most of the gliadins were solubilized and extracted at pH 5.3. The relative proportion of total polymeric proteins and unextractable polymeric proteins increased in subsequent fractions (Table 4), with the highest values in the pH 3.1 residue. SE-HPLC analyses of individual fractions (Fig. 4) also show the subsequent shift in MWD of the fractions to higher MW. This is clear in Table 5, where addition of these fractions to the base ours causes similar variations in percentage of polymeric protein greater than 250,000 Da. The solubility of gluten proteins depends on their molecular weight (Singh and MacRitchie, 2001). As the pH is lowered, relatively larger sized gluten proteins become soluble, and can be separated into different fractions by stepwise reduction of pH (Gupta et al., 1990; MacRitchie, 1985). 3.2.2. Reconstitution studies Extracted fractions were added to defatted base ours (Jagger and Soft) at a level of 1% (dry protein) on a total our weight basis. MacRitchie (1987) found signicant rheological differences at these levels of addition of protein fractions. The reconstituted ours were analyzed for their MWD, dough mixing properties, baking performance and biaxial extensional rheology. 3.2.2.1. SE-HPLC and MALLS (SEC-MALLS) analysis. Cut-off analysis, as described elsewhere (Bangur et al., 1997), of SE-HPLC chromatograms overlaid with MALLS signal, of all Jagger and Soft wheat ours (base and reconstituted) at 4 min intervals, helped to determine the elution time at which the polymers with MW of 250,000 Da were being eluted. The chromatograms were integrated to determine the monomeric (mainly gliadins) to polymeric (mainly glutenins) ratio, and relative proportion of polymeric proteins greater than 250,000 Da (Table 5). Addition of the gliadin rich (pH 5.3) fraction caused a decrease in the relative proportion of polymeric proteins greater than 250,000 Da. However, with addition of later fractions (pH4.9 to pH 3.1 residue), progressive increase in the proportion of polymeric proteins greater than 250,000 Da was observed. Use of the relative proportion of polymeric proteins greater than MW of 250,000 is based on a previous estimate by Bangur et al. (1997), which also seems to hold true in this study and is explained in Section 3.2.2.3.
Table 4 Yield, moisture and protein content of gluten protein fractions extracted by pH fractionation, and their SE-HPLC relative composition of polymeric proteins (all numbers in percentages) Fractions Yield Moisture (%) content (% wet basis) 4.5 0.08 4.6 0.16 4.1 0.15 4.2 0.37 4.3 0.09 4.9 0.26 Protein content (%) (14% m.b.) 83.8 0.05 78.0 0.08 79.6 0.12 78.5 0.07 73.2 0.14 45.8 0.18 Area (%) under chromatogram curve TPP UPP
Fig. 3. Cumulative percentage of total polymeric protein extracted as a function of nal pH of supernatant of Jagger gluten.
Elution of proteins through a SE-HPLC column is based on their hydrodynamic radii. Proteins with similar MW may vary in the conformation and thus the hydrodynamic radii in solution. Such proteins will not elute at the same time through the SE-HPLC column. An overlaid MALLS signal provides a true measure of MW at different elution times (Bean and Lookhart, 2001), providing nearly precise estimation of the relative proportion of polymeric proteins greater than 250,000 Da. 3.2.2.2. Dough mixing properties. To observe variations in mixing properties of the ours on addition of fractions, Mixograph analysis was performed with constant water absorption; i.e. 63% for Jagger wheat our and 60% for Soft wheat our. Mixograph dough development time (MDDT) was recorded from Mixograph traces of Jagger and Soft wheat ours. MDDT is plotted for each fraction as a function of the total protein that has been extracted (Fig. 5). Addition of the rst 35% (approximately) of total polymeric protein extracted from Jagger gluten (corresponding to the pH 5.3 fraction) caused a decrease in mixing requirements. In contrast, addition of subsequent fractions caused an increase in mixing requirements. These trends for MDDTs can be explained on the basis of SECMALLS data (Table 5). As expected, the decrease in the monomeric to polymeric ratio and increase in relative proportion of polymeric proteins greater than 250,000 Da lead to an increase in MDDT and corresponding decrease in breakdown. At a molecular level, mixing
pH 5.3 50.4 pH 4.9 20.5 pH 4.1 10.6 pH 3.5 4.2 pH 3.1 3.5 supernatant pH 3.1 residue 10.8
31.6 1.59 f 23.9 0.13 d 52.6 0.08 e 72.6 1.48 c 64.1 0.10 d 78.3 2.85 b 65.5 0.01 c 77.6 2.31 b 68.0 0.21 b 83.2 0.93 a 69.5 0.49 a 86.6 0.08 a Fig. 4. SE-HPLC chromatograms of total protein of Jagger gluten protein fraction. As pH is reduced, the percentage of polymeric protein increases and MWD shifts towards higher molecular weight.
TPPdtotal polymeric protein; UPPdunextractable polymeric protein Values represent mean standard deviation for duplicate determinations. Means with the same letter within columns are not signicantly different (p > 0.05).
38
Table 5 Parameters calculated from SEC-MALLS chromatograms of Jagger wheat our with addition of gluten protein fractions, added at 1% (dry protein) level (on our weight basis) Fraction added Jagger wheat our MON/POL None (Control) pH 5.3 pH 4.9 pH 4.1 pH 3.5 pH 3.1 supernatant pH 3.1 residue
a
Soft wheat our MON/POLa 1.41 0.01 a 1.43 0.03 a 1.27 0.07 b 1.28 0.01 b 1.30 0.01 b 1.16 0.00 c 1.16 0.02 c % POL > 250,000b 76.8 0.07 d 46.7 0.10 f 68.1 0.22 e 99.7 0.14 b,c 100.0 0.00 a 96.1 0.04 c 100.0 0.00 a
% POL > 250,000b 65.9 0.01 d 35.0 0.05 f 66.9 0.18 c 58.0 0.04 e 67.1 0.02 c 85.1 0.12 b 100 0.00 a
1.47 0.00 a 1.48 0.01 a 1.36 0.01 b,c 1.34 0.00 c,d 1.37 0.00 b 1.33 0.01 d 1.32 0.00 d
Values represent mean standard deviation for duplicate determinations. Means with the same letter within columns are not signicantly different (p > 0.05). a Monomeric to polymeric protein ratio. b Percentage of polymeric proteins ! 250,000 Da.
Fig. 6. Effect on loaf volumes (d) and strain hardening index (- - -) of Jagger (:) and Soft (-) wheat ours on addition of gluten protein fractions at a 1% (dry protein) level (on our weight basis).
is characterized by extension (stretching) of glutenin polymers and entangling of these stretched polymers (MacRitchie, 1986). This is achieved by work input and mixing intensity that must be above certain minimum critical levels (Kilborn and Tipples, 1972; Tipples and Kilborn, 1975). At a given work input and intensity of mixing, MDDT probably relates to the extension and entangling of the largest glutenin molecules (Singh and MacRitchie, 2001). The work input and mixing intensity of the our is therefore a function of its MWD. It has been observed that when the percentage of UPP increases and/or the MWD of the polymeric fraction is shifted to higher MWs, the MDDT increases, as requirements for work input increase (Gupta et al., 1993). 3.2.2.3. Baking performance. Loaf volumes (Fig. 6) decreased on addition of the pH 5.3 fraction. However, on addition of subsequent fractions that lead to increase in MDDT, increase in loaf volume was observed, which after reaching a maximum, again decreased. This maximum was achieved in the case of Jagger on addition of total polymeric proteins that were extracted after the rst 60% and before the rst 80% corresponding to the pH 4.1 fraction. In the case of Soft, the maximum occurred on addition of total polymeric proteins that were extracted after the rst 35% and before the rst 70%, corresponding to the pH 4.9 fraction (Figs. 3 and 6). It seems that the optimum balance of strength and extensibility was achieved on addition of these fractions to base ours, thus giving best baking performance. Bangur et al. (1997) reported that the ratio of polymeric proteins with MWs above the threshold molecular weight (MT) to those (i.e.
polymeric proteins) with MWs below MT was approximately 60:40. The MT for polymeric proteins was estimated to be 250,000 Da. Polymeric proteins greater in molecular weight than 250,000 Da confer strength to the entangled gluten protein network and those with molecular weight less than this may counter the strength by acting as diluents (or plasticizers). This 60:40 proportion is nearly achieved on addition of the pH 4.1 fraction in Jagger wheat our and pH 4.9 fraction in Soft wheat our (Table 5) giving maximum loaf volumes and strain hardening index. Any shift in the proportion shows deleterious effects on baking performance (Fig. 6). 3.2.2.4. Biaxial extensional rheology. Signicant differences (P < 0.0001) in biaxial extensional rheological parameters (strain hardening index and failure strain) were recorded for addition of fractions to the base ours. Variations in strain hardening index paralleled those of loaf volume, for both ours (Fig. 6). Analogous variations in strain hardening index and loaf volume show that this parameter is a good determinant of rheological requirements for the glutenstarch matrix in terms of bread making. As reported in some previous studies (Dobraszczyk and Roberts, 1994; Dobraszczyk et al., 2003; Tronsmo et al., 2003), in this case also, strain hardening shows a high degree of correlation with loaf volume and failure strain (Fig. 7). The Mixograph parameter (MDDT), on the other hand, correlates well with MWD but does not predict baking performance equally well. The concept of strain hardening explains the requirements for maximum inatability of gas cells in bread making without failure, to give good loaf volume and crumb grain. In a glutenstarch matrix, gluten proteins form a continuous network and the rheology of the dough is that of a continuous gluten protein network. It is only the fraction of glutenins with MW greater than MT that confers strength, as suggested by Bangur et al. (1997), whereas smaller ones act as diluents, preventing additional physical constraints or entanglements and enhancing extensibility (Bersted and Anderson, 1990; Singh and MacRitchie, 2001). To get maximum inatability of the gas cells, the glutenstarch matrix around them must stretch to its maximum extensibility without breaking, i.e. the glutenins/polymeric proteins (mainly with MW > MT; i.e. 250,000 Da) must be stretched to their maximum length through entanglements, as described by entanglement network theory (MacRitchie and Laandra, 1997; Singh and MacRitchie, 2001). Addition of the gliadin rich fraction, i.e. the pH 5.3 fraction (Fig. 4), signicantly increases diluent concentration that will lead to disentanglement of the gluten protein network without sufcient elongation when subjected to extensional forces, thus leading to earlier failure of the gas cells. As MWD shifts
Fig. 5. Effect on Mixograph peak times of Jagger (- - -) and Soft (d) wheat ours on addition of gluten protein fractions at a 1% (dry protein) level (on our weight basis).
39
Fig. 7. Scatter plots of strain hardening index vs. failure strain and loaf volume (cc) for different Jagger and Soft wheat our doughs exhibiting high degree of correlation.
towards larger glutenins, the strength increases (Gupta et al., 1993). On doing so, it passes through maximum extensibility or strain (Hencky), beyond which the balance shifts towards strength and the gluten protein network is no longer as extensible. This leads to decrease in strain hardening index as well as loaf volume upon addition of the latter extracted fractions (Fig. 6).
4. Conclusion Results of this study provide clear evidence for the presence of liquid lamellae, thus supporting the dual lm hypothesis. The liquid lamellae act as a secondary stabilizing mechanism and is on the inner side of the glutenstarch matrix enveloping the gas cells (Gan et al., 1995). Stability of the glutenstarch matrix, which is the primary stabilizing factor for expanding gas cells against disproportionation and coalescence, depends on its tendency to strain harden. The phenomenon of strain hardening appears to depend on the balance between strength and extensibility of the entangled network of polymeric proteins of wheat our. Extensibility ensures slippage of the maximum number of statistical segments between entanglements (Singh and MacRitchie, 2001), whereas strength prevents disruption of the entangled network of polymeric proteins. Thus, to ensure stability of gas cells, the dough needs to be sufciently extensible to respond to gas pressure but also strong enough to resist collapse. Differences in gluten quality, as demonstrated in this study can signicantly affect the bread making potential. Strength is conferred by the fraction of polymeric proteins having molecular weight greater or equivalent to MT (250,000), and the fraction of gluten protein smaller than MT may counter the strength by acting as diluents. The optimum balance seems to exist when the relative proportions of polymeric proteins greater and smaller than MT are roughly 60:40. Shift in the balance to either side will decrease loaf volume. Increase in smaller proteins (less than MT) may decrease stability of the glutenstarch matrix due to a lesser number of entanglements per chain. On the other hand, increase in strength conferring proteins may prevent sufcient expansion of the glutenstarch matrix required to increase loaf volume due to reduced slippage of gluten polymers through entanglement nodes as a result of increase in number of entanglements per chain.
The secondary stabilizing mechanism involves thin liquid lamellae stabilized by adsorbed surface active compounds (lipids and proteins) at the gasliquid interface. Liquid lamellae prevent coalescence and disproportionation of gas cells when they come in close contact with each other during the late proving and early baking stages of bread making i.e. when discontinuities begin to appear in the glutenstarch matrix. The study demonstrates that the our lipids at their natural levels do not inuence rheological properties of the glutenstarch matrix surrounding the gas cells, as measured by the dough ination system. Nevertheless, the small amounts in which these lipids are naturally present are sufcient to inuence surface properties. The exact mechanism by which these lipids and other surface active components (proteins) stabilize liquid lamellae is discussed in the succeeding paper (Part II).
References
AACC, 2001. Approved Methods of the AACC, tenth ed. American Association of Cereal Chemists, St. Paul, Minnesota, USA. Baker, J.C., Mize, M.D., 1946. Gas occlusion during dough mixing. Cereal Chemistry 23, 3951. Bangur, R., Batey, I.L., McKenzie, E., MacRitchie, F., 1997. Dependence of extensograph parameters on wheat protein composition measured by SE-HPLC. Journal of Cereal Science 25, 237241. Batey, I.L., Gupta, R.B., MacRitchie, F., 1991. Use of size-exclusion high-performance liquid chromatography in the study of wheat our proteins: an improved chromatographic procedure. Cereal Chemistry 68, 207209. Bean, S.R., Lookhart, G.L., 2001. Factors inuencing the characterization of gluten proteins by size-exclusion chromatography and multiangle laser light scattering (SEC-MALLS). Cereal Chemistry 78, 608618. Bersted, B.H., Anderson, T.G., 1990. Inuence of molecular weight and molecular weight distribution on the tensile properties of amorphous polymers. Journal of Applied Polymer Science 39, 499514. Bloksma, A.H., 1990. Dough structure, rheology, and baking quality. Cereal Foods World 35, 237244. Chin, N.L., Campbell, G.M., 2005. Dough aeration and rheology. II. Effects of our type, mixing speed and total work input on aeration and rheology of bread dough. Journal of the Science of Food and Agriculture 85, 21942202. DeStefanis, V.A., Ponte Jr., J.G., 1976. Studies on breadmaking properties of wheatour nonpolar lipids. Cereal Chemistry 53, 636642. Dobraszczyk, B.J., 1997. Development of a new dough ination system to evaluate doughs. Cereal Foods World 42, 516519. Dobraszczyk, B.J., Roberts, C.A., 1994. Strain hardening and dough gas cell-wall failure in biaxial extension. Journal of Cereal Science 20, 265274. Dobraszczyk, B.J., Salmanowicz, B.P., 2008. Comparison and predictions of baking volume using large deformation rheological properties. Journal of Cereal Science 47, 292301.
40
B.S. Sroan et al. / Journal of Cereal Science 49 (2009) 3240 MacRitchie, F., 1987. Evaluation of contributions from wheat protein fractions to dough mixing and breadmaking. Journal of Cereal Science 6, 259268. MacRitchie, F., Gras, P.W., 1973. The role of our lipids in baking. Cereal Chemistry 50, 292302. MacRitchie, F., Kasarda, D.D., Kuzmicky, D.D., 1991. Characterization of wheat protein fractions differing in contributions to breadmaking quality. Cereal Chemistry 68, 122130. MacRitchie, F., Laandra, D., 1997. Structurefunctionality relationships of wheat proteins. In: Damodaran, S., Paraf, A. (Eds.), Food Proteins and Their Applications. Marcel Dekker, Inc., New York, USA, pp. 293324. McCormack, G., Panozzo, J., MacRitchie, F., 1991. Contributions to breadmaking of inherent variations in lipid content and composition of wheat cultivars. II. Fractionation and reconstitution studies. Journal of Cereal Science 13, 263274. Mills, E.N.C., Wilde, P.J., Salt, L.J., Skeggs, P., 2003. Bubble formation and stabilization in bread dough. Food and Bioproducts Processing 81, 189193. Paternotte, T.A., Orsel, R., Hamer, R.J., 1993. Interactions between our proteins and our lipids at the liquid/air interface. In: Gluten Proteins. Association of Cereal Research, Detmold, Germany, pp. 207217. Ross, A., MacRitchie, F., 1995. Interactions of wheat proteins, carbohydrates and lipids. In: Gaonkar, A.G. (Ed.), Ingredient Interactions Effects on Food Quality. Marcel Deker, New York, USA, pp. 321356. Singh, N.K., Donovan, G.R., Batey, I.L., MacRitchie, F., 1990. Use of sonication and size-exclusion high-performance liquid chromatography in the study of wheat our proteins. I. Dissolution of total proteins in the absence of reducing agents. Cereal Chemistry 67, 150161. Singh, H., MacRitchie, F., 2001. Application of polymer science to properties of gluten. Journal of Cereal Science 33, 231243. Termonia, Y., Smith, P., 1988. Kinetic model for tensile deformation of polymers. 2. Effect of entanglement spacing. Macromolecules 21, 21842189. Tronsmo, K.M., Frgestad, E.M., Schoeld, J.D., Magnus, E.M., 2003. Wheat protein quality in relation to baking performance evaluated by the Chorleywood bread process and a hearth bread baking test. Journal of Cereal Science 38, 205215. Tipples, K.H., Kilborn, R.H., 1975. Unmixingdthe disorientation of developed bread doughs by slow speed mixing. Cereal Chemistry 52, 248262. van Vliet, T., Janssen, A.M., Bloksma, A.H., Walstra, P., 1992. Strain hardening of dough as a requirement for gas retention. Journal of Texture Studies 23, 439460.
Dobraszczyk, B.J., Smewing, J., Albertini, M., Maesmans, G., Schoeld, J.D., 2003. Extensional rheology and stability of gas cell walls in bread doughs at elevated temperatures in relation to breadmaking performance. Cereal Chemistry 80, 218224. Gan, Z., Angold, R.E., Williams, M.R., Ellis, P.R., Vaughan, J.G., Galliard, T., 1990. The microstructure of gas retention of bread dough. Journal of Cereal Science 12, 1524. Gan, Z., Ellis, P.R., Schoeld, J.D., 1995. Gas cell stabilization and gas retention in wheat bread dough. Journal of Cereal Science 21, 215230. Gandikota, S., MacRitchie, F., 2005. Expansion capacity of doughs: methodology and applications. Journal of Cereal Science 42, 157163. Gupta, R.B., Bekes, F., Gras, P.W., MacRitchie, F., 1990. Functionality of glutenin, gliadin and secalin fractions as measured by extensograph and mixograph. In: Gluten Proteins. Association of Cereal Research, Detmold, pp 550559. Gupta, R.B., Khan, K., MacRitchie, F., 1993. Biochemical basis of our properties in bread wheats. I. Effects of variation in the quality and size distribution of polymeric protein. Journal of Cereal Science 18, 2341. Hoseney, R.C., 1992. Physical chemistry of bread dough. In: Schwartzberg, H.G., Hartel, R.W. (Eds.), Physical Chemistry of Foods. Marcel Dekker, Inc., New York, USA, pp. 443457. Kilborn, R.H., Tipples, K.H., 1972. Factors affecting mechanical dough development. I. Effect of mixing intensity and work input. Cereal Chemistry 49, 3447. Lundh, G., MacRitchie, F., 1989. Size exclusion HPLC characterization of gluten protein fractions varying in breadmaking potential. Journal of Cereal Science 10, 247253. MacRitchie, F., 1976a. The liquid phase of dough and its role in baking. Cereal Chemistry 53, 318326. MacRitchie, F., 1976b. Monolayer compression barrier in emulsion and foam stability. Journal of Colloid and Interface Science 56, 5356. MacRitchie, F., 1978. Differences in baking quality between wheat ours. Journal of Food Technology 13, 187194. MacRitchie, F., 1985. Studies of the methodology for fractionation and reconstitution of wheat ours. Journal of Cereal Science 3, 221230. MacRitchie, F., 1986. Physicochemical processes in mixing. In: Blanshard, J.M.V., Frazier, P.J., Galliard, T. (Eds.), Chemistry and Physic of Baking. Royal Society of Chemistry, London, UK, pp. 132146.

Mechanism of gas cell stabilization in breadmaking. II. The secondary liquid lamellae
Baninder S. Sroan*, Finlay MacRitchie
Department of Grain Science and Industry, Kansas State University, Manhattan, KS 66506, USA
Article history: Received 19 December 2007 Received in revised form 28 June 2008 Accepted 9 July 2008 Keywords: Breadmaking Gas cell stability Liquid lamella Glutenstarch matrix Coalescence Disproportionation
a b s t r a c t
The study for the rst time demonstrates that our lipids at their natural levels do not affect dough rheology as measured by bubble ination, thus indicating the presence of liquid lamellae as an independent secondary gas cell stabilizing mechanism in bread dough. The liquid lamellae, stabilized by adsorbed surface active compounds, plays its role during the later proving and early baking stage, when discontinuities occur in the glutenstarch matrix surrounding gas bubbles. To study this secondary stabilizing mechanism, different lipid fractions were added incrementally to the defatted ours. No effects were observed on the rheological properties of the dough. However, large effects on the loaf volume were measured. The additives used were the total our lipid and its polar and non-polar fractions and the fatty acids palmitic, linoleic and myristic. Polar lipids and palmitic acid had positive or little effect on loaf volume, respectively. Non-polar lipid, linoleic and myristic acids had negative effects on loaf volume. The different effects of the lipid fractions are thought to be related to the type of monolayer that is formed. Polar lipid and palmitic acid form condensed monolayers at the air/water interface whereas non-polar lipid, linoleic and myristic acids form expanded monolayers. 2008 Elsevier Ltd. All rights reserved.
1. Introduction Concentration of uniformly sized gas bubbles in the liquid phase of dough during mixing is essential for their sufcient expansion without failure during the breadmaking process for good loaf volume and crumb grain. There is evidence that the liquid lamellae surrounding expanding gas cells act as a secondary protection together with the primary glutenstarch matrix, thus preventing their coalescence and disproportionation as outlined in Part-I. The secondary stabilizing mechanism of thin liquid lamellae (lms) operates when gas bubbles come in close contact during later proving and early baking when discontinuities begin to appear in the glutenstarch matrix. Stability of the liquid lamellae depends on surface active compounds (proteins and lipids) (Baker et al., 1946; MacRitchie, 1976a; Mills et al., 2003), that are adsorbed at the gasliquid interface. Analogous variations in foaming properties of the dough liquor, and bread loaf volume and crumb structure (MacRitchie, 1976a) provide plausible evidence of surface action of these compounds at the gasliquid interface. Contributions of liquid lm stability to dough expansion can be assessed from different studies. It has been shown that our lipids, though present in very small amounts (w11.5%) have signicant effects on
* Corresponding author. Present address: FritoLay, 7701 Legacy Dr., 3T-124, Plano, TX 75024, USA. Tel.: 1 972 334 4526; fax: 1 972 334 2329. E-mail address: baninder@k-state.edu (B.S. Sroan). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.07.004
baking performance in terms of volume and crumb structure (MacRitchie and Gras, 1973; Mills et al., 2003). Polar lipids improve bread loaf volume and crumb grain, whereas non-polar lipids have the opposite effects (MacRitchie and Gras, 1973; Ponte and DeStefanis, 1969). DeStefanis and Ponte (1976) demonstrated that an unsaturated free fatty acid (linoleic acid), a component of the nonpolar lipid fraction is responsible for detrimental effects on loaf volume. Mills et al. (2003) reviewed some of the more recent studies on effects of surface active components of wheat our on breadmaking performance. The most probable mechanism by which these surface active components control stability of gas cells is through formation of monolayers at the gasliquid interface (Mills et al., 2003; Ross and MacRitchie, 1995). Saturated free fatty acids, like palmitic and stearic, form condensed monolayers, which are highly incompressible with high surface elasticity and these molecules do not desorb easily from the monolayers. These condensed monolayers generate substantial elastic restoring forces, which resist destabilization of liquid lamellae when changes in interfacial area occur (MacRitchie, 1976b). Molecules like digalactosyl diglyceride (DGDG) with strong polar head groups are expected to orient themselves strongly at the interface as condensed monolayers. Unsaturated free fatty acids like linoleic acid give expanded monolayers, which are relatively compressible and slightly soluble in the subphase, leading to instability of the liquid lamellae (MacRitchie, 1976b). Rich in surface active compounds (lipids,
42
B.S. Sroan, F. MacRitchie / Journal of Cereal Science 49 (2009) 4146
proteins, arabinoxylans, added surfactants, etc.), the liquid phase of dough is a very competitive environment in terms of adsorption of these compounds at the gasliquid interface (Mills et al., 2003; Ross and MacRitchie, 1995). Proteins (soluble or insoluble) in wheat our are capable of forming condensed monolayers. Pure protein monolayers possess high surface elasticity. When lipids are introduced in the system, surface elasticity drops as lipids compete with proteins for the gasliquid interface (Paternotte et al., 1993; Salt et al., 2006). It seems that the type of monolayer formed will either stabilize or destabilize the freshly occluded gas cells, and any changes in monolayer state will inuence the gas cell stability at all stages of breadmaking. Therefore, the stability of liquid lamellae determines the ease with which gas cells are concentrated during mixing and ensure their permanence during the entire breadmaking process. The objective of this part of the study was to understand the possible mechanism by which liquid lamellae are stabilized. This part of the study provides further evidence in support of independent existence of liquid lamellae as a secondary stabilizing mechanism supporting expanding gas cells. It shows for the rst time that lipids, at their natural levels, have no effect on the rheological properties of the starch/protein (glutenstarch) matrix as measured by bubble ination and their effects are therefore independent of the stabilizing mechanism discussed in Part-I. 2. Experimental Materials and various procedures (analytical, dough mixing properties, test baking, image analysis, biaxial extensional rheology and statistical analysis) used in the study are explained in detail in Part-I. 2.1. Free fatty acids Palmitic acid (99%), linoleic acid (60%), and myristic acid (99%) were purchased from Sigma-Aldrich, USA. Values in brackets are of minimum purity by gas chromatography as per certicate of analysis by manufacturer. Linolenic acid was the main impurity in linoleic acid. 2.2. Lipid extraction from our Natural our lipids were extracted using three batch extractions with chloroform in a glass beaker, followed by Buchner ltration through Whatman No. 1 lter paper (MacRitchie and Gras, 1973). 200 g of our and 400 ml of chloroform were used for each extraction. The defatted our was spread out on a at glass tray in a fume hood for 12 h to allow evaporation of solvent. 2.3. Lipid fractionation Natural lipids of both ours in chloroform were collected together and concentrated under vacuum using a rotary evaporator at a temperature <30 C. The combined our lipids were then separated into polar and non-polar fractions by the method of Ponte and DeStefanis (1969). High purity grade silica gel without binder (Sigma-Aldrich, USA) of average particle size 525 mm was used in the procedure. The percentage of natural our lipid extracted with chloroform was 0.91% (our weight basis). The percentage of polar and non-polar fractions separated out was 0.36% and 0.55% (our weight basis), respectively. These fractions i.e. natural our lipids, polar fraction and non-polar fraction, were analyzed for their composition, qualitatively, using the thin-layer chromatography (TLC) method of McCormack et al. (1991). TLC patterns (Fig. 1) show efcient separation of natural our lipids into polar and non-polar fractions. After extraction and analysis, the
Fig. 1. Thin-layer chromatogram of (1) natural our lipid and its (2) polar and (3) nonpolar fractions.
fractions were solubilized in chloroform and nitrogen was bubbled through the liquids to remove any solubilized oxygen. The liquids (fraction solvent) were then refrigerated under nitrogen atmosphere until their use. For use in baking and rheology measurements, these were again concentrated under vacuum using a rotary evaporator. 2.4. Addition of lipids to defatted our Different lipid types (polar, non-polar) and free fatty acids (linoleic acid, palmitic acid and myristic acid) were added to the defatted our in varying amounts. For incorporation of intact natural our lipids into defatted our, the parent our was mixed with defatted our in different proportions, to give different our lipid levels. All other lipid additions were made as a percentage of total our lipids; for example, addition of 60% non-polar fraction means 60% of 0.91 g was added to 100 g of defatted our. 3. Results and discussion To gain understanding of the mechanism by which surface active components (lipids and proteins) of wheat our dough affect stability of the liquid lamellae surrounding expanding gas cells, baking trials were conducted by adding different lipid types (natural our lipid and its polar and non-polar factions) and free fatty acids (linoleic, palmitic and myristic acids) to defatted ours at different levels of addition (on original our weight basis). Effect of lipids was also tested on biaxial extensional rheology of the doughs. 3.1. Effect of variations in lipid types and free fatty acids, and their levels on baking performance 3.1.1. Breadmaking Loaf volume trends observed on incremental addition of different lipid types and free fatty acids were mainly similar for both Jagger and soft wheat ours, except for some variations in certain cases (Figs. 2 and 3). Within each lipid type and free fatty acid, signicant differences were observed in response to varying levels, except palmitic acid in the case of Jagger our. Incremental addition of natural our lipids back into parent ours caused loaf volume to decrease, and, after reaching a minimum, to increase. Incremental addition of the polar lipids also displayed initial
43
Fig. 2. Loaf volume vs. different lipid types, added to defatted Jagger wheat our as percentage of natural our lipids.
depression for small additions followed by increase in loaf volume. This increase in volume was appreciably larger than what was observed for the same levels of natural our lipids. Addition of increasing levels of the non-polar lipids and linoleic acid caused a continuous decrease in loaf volume for both the ours. Incremental addition of palmitic acid produced no change in loaf volume of Jagger our and slight depression in soft wheat our. Incremental addition of myristic acid caused depression in loaf volume at initial levels, however, at considerably higher levels of addition, some reversal was observed. This increase, in the case of myristic acid, leveled off at the volumes equivalent to those of
Fig. 3. Loaf volume vs. different lipid types, added to defatted soft wheat our as percentage of natural our lipids.
intact ours (100% of natural our lipids). Maximum addition of myristic acid in Jagger was 400% of natural our lipid content. This was a very high level of addition in comparison to other lipid types and free fatty acids used in the study; this was done to have a clear picture of the loaf volume trend in this case. The results for natural our lipids and their polar and non-polar fractions agreed qualitatively with those of MacRitchie and coworkers (McCormack et al., 1991; MacRitchie, 1977; MacRitchie, 1978; MacRitchie and Gras, 1973), and those of unsaturated (linoleic) and saturated (palmitic) free fatty acids concurred with results of DeStefanis and Ponte (1976). Effects of different lipid types on loaf volume and crumb structure correlated with results from lm balance studies reported by various workers (Gaines, 1966; MacRitchie, 1990). This substantiated our earlier results (Part-I), providing evidence for the presence of liquid lamellae around expanding gas cells. The results also indicated that the most probable mechanism by which different lipid types inuenced the gas cell stability was through their adsorption as monolayers at the gasliquid interface. The monolayers formed are either condensed or expanded (Gaines, 1966; MacRitchie, 1990). Condensed monolayers are characterized by close packing of surface active molecules, with polar head groups oriented towards water and non-polar hydrocarbon chains towards air. Expanded monolayers are formed due to loose packing of surface active molecules, occupying a much higher area per molecule. Condensed monolayers are relatively incompressible and elastic compared to expanded ones and are less easily desorbed than expanded monolayers (MacRitchie, 1976b). On expansion of gas cells, condensed monolayers provide elastic restoring forces, contributing to resistance to collapse of liquid lamellae with change of interfacial area (MacRitchie, 1976b). In defatted ours, proteins are the only surface active components forming monolayers. This results in a higher loaf volume and a ner and more uniform crumb structure as proteins provide quite stable monolayers and do not desorb easily from the gasliquid interface. This has been attributed to suitable congurations adopted by protein molecules at the interface and to their large sizes (Larsson et al., 2006; MacRitchie, 1990). Addition of different lipid types may lead to the formation of mixed monolayers (Mills et al., 2003; Ross and MacRitchie, 1995; Salt et al., 2006). The occurrence of minima in loaf volume on addition of natural our lipid and its polar fraction is analogous to the decrease in stability of mixed lms reported by Paternotte et al. (1993) on passage from a pure protein lm to a lipid (monogalactosyl monoglyceride) protein stabilized lm. Variations in percentage of natural our lipids at which the minimum volume is reached could be attributed to differences in protein composition of the two ours. A study by Salt et al. (2006) also indicated that mixed proteinlipid interfaces are relatively less elastic and therefore more unstable compared to pure protein ones, which possess high surface elasticity. Palmitic acid, a saturated free fatty acid, and polar lipids having digalactosyl diglyceride (DGDG), due to their tendency to form condensed monolayers (Gaines, 1966; MacRitchie, 1990), had either no effect or positive effect on loaf volumes, respectively. On the other hand, the probable formation of an expanded monolayer by linoleic acid resulted in detrimental effects on loaf volume. Myristic acid is a saturated free fatty acid but, due to its shorter chain length, forms expanded monolayers (MacRitchie, 1990), thus decreasing loaf volume. At higher levels of myristic acid addition, a reversal of the effect is observed (Figs. 2 and 3). This could be attributed to the presence of impurities in myristic acid, which is w99% pure by gas chromatographic analysis. However, such purity analysis does not indicate surface chemical purity of surface active compounds like myristic acid (MacRitchie, 1990). The presence of impurities, even in minor amounts, can inuence surface properties. This might have been the case at higher levels of addition of myristic acid in
44
Table 1 Crumb structure responses of Jagger and soft wheat our breads to different lipid types and levels, added to defatted ours as percentage of natural our lipids % Of natural our lipids Jagger wheat our Number of cells Natural our lipids 0% 20% 40% 50% 60% 80% 100% Flour polar lipids 0% 20% 40% 66% 132% 200% Flour non-polar lipids 0% 15% 30% 60% 132% 200% Linoleic acid 0% 66% 132% 200% 250% Palmitic acid 0% 66% 132% 200% 250% Myristic acid 0% 66% 132% 200% 250% 330% 400% 2000.0 11.3 a 1935.0 222.03 a,b 1880.5 78.5 a,b 1898.0 59.4 a,b 1788.5 81.3 a,b 1907.5 2.1 a,b 1749.0 41.0 b 2361.5 7.8 a 2307.0 157.0 a 2001.0 24.0 b 1985.0 0.0 b 2285.0 0.0 a 2385.0 0.0 a 2436.0 113.1 a 2306.5 126.6 a 1989.0 69.3 b,c 2202.5 187.4 a,b 1937.5 47.4 b,c 1761.5 0.0 c 2452.5 26.2 a 1803.0 220.6 b 1796.5 67.2 b 1473.5 169.0 b 1505.5 13.4 b 2477.5 139.3 a 2438.5 103.9 a 2522.5 171.8 a 2331.0 14.1 a 2285.5 126.5 a 2328.0 63.6 a 2368.0 190.9 a 2404.5 54.5 a 2447.5 150.6 a 2244.5 17.7 a 2324.0 84.9 a 2632.0 35.4 a Average cell elongation (C-cell score) 1.58 0.01 b 1.59 0.05 b 1.64 0.06 a,b 1.68 0.04 a 1.67 0.05 a,b 1.62 0.01 a,b 1.63 0.01 a,b 1.56 0.03 c,d 1.57 0.01 b,c,d 1.60 0.04 a,b,c 1.62 0.00 a 1.54 0.00 d 1.61 0.00 a,b 1.57 0.01 b 1.57 0.00 b 1.60 0.04 b 1.61 0.01 a,b 1.65 0.01 a 1.61 0.00 a,b 1.63 0.06 a 1.68 0.02 a 1.72 0.04 a 1.70 0.04 a 1.72 0.00 a 1.61 0.04 a 1.64 0.02 a 1.63 0.06 a 1.58 0.00 a 1.59 0.01 a 1.62 0.01 a 1.62 0.02 a 1.61 0.01 a 1.64 0.00 a 1.65 0.00 a 1.66 0.00 a 1.65 0.00 a Soft wheat our Number of cells 2544.5 187.4 a,b 2521.0 110.3 a,b,c 2210.5 115.3 d 2209.0 79.2 d 2347.5 2.1 b,c,d 2196.0 86.3 d 2229.0 114.6 d 2107.0 26.9 b 1876.5 37.5 c 2389.0 0.0 a 2183.0 0.0 b 2180.5 23.3 b 2092.0 216.4 b 2110.0 22.6 a 2036.5 95.5 a,b 1832.0 144.2 b 1950.0 104.7 a,b 2053.0 14.1 a,b 2060.5 29.0 a,b 2290.0 158.4 a 1926.0 116.0 b 2019.0 144.2 a,b 1994.0 120.2 a,b 2060.5 29.0 a 2043.0 69.3 a 1940.5 68.6 a 1880.5 17.7 a 1829.5 204.4 a 2100.0 135.8 a 1954.0 24.0 a 2263.5 54.5 a 2270.5 147.8 a 2156.5 188.8 a Average cell elongation (C-cell score) 1.69 0.01 a,b 1.68 0.01 a,b,c 1.65 0.03 b,c 1.65 0.02 c 1.70 0.02 a 1.66 0.00 a,b,c 1.69 0.03 a,b 1.60 0.02 c 1.66 0.02 a,b 1.64 0.00 b 1.67 0.00 a,b 1.69 0.03 a 1.67 0.01 a,b 1.60 0.02 a 1.59 0.01 a 1.68 0.05 a 1.66 0.05 a 1.64 0.01 a 1.60 0.00 c 1.68 0.02 b 1.69 0.02 a,b 1.73 0.01 a 1.71 0.02 a 1.60 0.00 b 1.64 0.01 a,b 1.66 0.06 a,b 1.72 0.04 a 1.65 0.03 b 1.68 0.01 1.68 0.01 1.68 0.01 1.68 0.01 1.68 0.01 b b b b b
Fig. 4. C-cell images of Jagger wheat our bread slices showing differences in gas cell expansion at different polar lipid levels (represented as percentages) added to defatted our as percentage of natural our lipids.
45
Table 2 Mean bubble ination rheological responses of Jagger wheat our doughs to different lipid types and levels, added to defatted ours as percentage of natural our lipids % Of natural our lipid Natural our lipids 0% 60% 100% Flour polar lipids 40% 200% Max. stress (kPa) 571.62 305.45 a 491.68 237.21 a 598.62 112.05 a 480.85 67.26 a 660.22 130.74 a Failure strain (Hencky) 2.58 0.18 a 2.60 0.12 a 2.66 0.04 a 2.56 0.04 a 2.64 0.05 a 2.74 0.16 a 2.66 0.04 a 2.62 0.12 a 2.65 0.03 a 2.63 0.10 Strain hardening index 2.21 0.12 b,c 2.38 0.064 a,b 2.44 0.06 a 2.18 0.03 c 2.22 0.04 b,c 2.44 0.09 a 2.36 0.10 a,b,c 2.23 0.09 b,c 2.32 0.10 b,c 2.31 0.12
Flour non-polar lipids 60% 936.93 560.68 a Linoleic acid 132% Palmitic acid 132% Myristic acid 250% Over all average 650.21 18.07 a 598.55 239.23 a 613.71 90.70 a 623.58 236.59
Fig. 5. C-cell images of Jagger wheat our bread slices showing differences in gas cell expansion at different non-polar lipid levels (represented as percentages) added to defatted our as percentage of natural our lipids.
baking studies. Therefore analysis of surface chemical purity (Lunkenheimer and Miller, 1987) of added surfactants, though not performed in this study, will be of great advantage in future studies. 3.1.2. Crumb structure Image analysis of bread crumb showed that addition of different lipid types and variations in their levels resulted in negligible differences in number of gas cells and average cell elongation (Table 1). These insignicant differences in the number of gas cells for a particular our when treated with different levels of the same lipid type and free fatty acid, suggest that higher loaf volumes were due to an increase in expansion capacity of gas cells and not due to their number. That is, the lm stabilizers (polar lipids and palmitic acid) allowed gas cells to expand, increasing the volume (Fig. 4). With improvement in expansion capacity, the size of gas cells increased and crumb structure changed from ne to more uniformly distributed larger gas cells. The inability of gas cells to expand (Fig. 5) also explains the earlier observations of relatively ne crumb appearance with addition of non-polar lipids reported by Ponte and DeStefanis (1969). However, an exception was observed in the case of linoleic acid addition to Jagger our. Incremental additions of linoleic acid to Jagger our caused small but continuous decrease in number of gas cells (Table 1). Linoleic acid might also be detrimental to initial concentration and stability of gas cells during mixing, thus hindering the ease with which gas cells are occluded into the liquid phase of dough. This initial adsorption depends upon the diffusion coefcient of surface active compounds (lipids and proteins) in the liquid phase of the dough (MacRitchie, 1990), and will vary from our to our. The nature of surface active compounds which has been adsorbed initially at the interface to form a monolayer will stabilize or destabilize freshly occluded gas cells, thus affecting the ease with which they are concentrated (Larsson et al., 2006). Negligible variations in average cell elongation indicated that different lipid types and free fatty acids at the levels of addition used might not be inuencing rheology of the glutenstarch matrix, since cell elongation is thought to be associated with dough rheological properties (Gandikota and MacRitchie, 2005). 3.2. Effect of variations in lipid types and free fatty acids, and their levels on biaxial extensional rheology of glutenstarch matrix Biaxial extensional rheological tests were performed on Jagger and soft doughs with different lipid types and their levels, in order
to investigate possible independent action of lipids on baking performance (loaf volume and crumb structure). Different lipid types and free fatty acids at the levels used did not inuence biaxial extensional rheological parameters (maximum stress, failure strain and strain hardening index) of the doughs (Tables 2 and 3). However, higher values were observed for Jagger than soft wheat our. Though minor differences were observed for strain hardening indices within a particular our at some treatments, no specic trend was observed, thus attributing the variation to slight experimental scatter. As for the case in Part-I of this study, higher values of biaxial extensional rheological parameters were observed in the case of Jagger our, indicating its superior breadmaking potential (Dobraszczyk et al., 2003). Inability of different lipid types and free fatty acids to cause any effect on biaxial extensional rheology, demonstrates that their action is independent of rheology of the
Table 3 Mean bubble ination rheological responses of soft wheat our doughs to different lipid types and levels, added to defatted ours as percentage of natural our lipids % Of natural our lipid Natural our lipids 0% 40% 100% Flour polar lipids 20% 200% Flour non-polar lipids 60% Linoleic acid 132% Palmitic acid 132% Myristic acid 250% Over all average Max. stress (kPa) 120.80 12.76 a 125.42 12.14 a 102.40 5.45 a 105.49 3.52 a 130.17 1.61 a 108.71 27.51 a 104.35 12.49 a 116.21 0.52 a 115.44 23.14 a 114.19 13.69 Failure strain (Hencky) 1.79 0.00 a 1.85 0.02 a 1.66 0.05 b 1.68 0.02 b 1.84 0.02 a 1.83 0.02 a 1.81 0.06 a 1.82 0.01 a 1.78 0.02 a 1.79 0.07 Strain hardening index 1.76 0.01 a 1.74 0.01 a 1.64 0.04 c 1.66 0.00 b,c 1.72 0.02 a,b 1.73 0.01 a,b 1.75 0.07 a 1.79 0.01 a 1.72 0.03 a 1.72 0.05
46
B.S. Sroan, F. MacRitchie / Journal of Cereal Science 49 (2009) 4146 Gandikota, S., MacRitchie, F., 2005. Expansion capacity of doughs: methodology and applications. Journal of Cereal Science 42, 157163. Larsson, K., Quinn, P., Sato, K., Tiberg, F., 2006. Lipids: Structure, Physical Properties and Functionality. The Oily Press, Bridgwater, England. Lunkenheimer, K., Miller, R., 1987. A criterion for judging the purity of adsorbed surfactant layers. Journal of Colloid Interface Science 120, 176183. McCormack, G., Panozzo, J., MacRitchie, F., 1991. Contributions to breadmaking of inherent variations in lipid content and composition of wheat cultivars. II. Fractionation and reconstitution studies. Journal of Cereal Science 13, 263274. MacRitchie, F., 1976a. The liquid phase of dough and its role in baking. Cereal Chemistry 53, 318326. MacRitchie, F., 1976b. Monolayer compression barrier in emulsion and foam stability. Journal of Colloid and Interface Science 56, 5356. MacRitchie, F., 1977. Flour lipids and their effects in baking. Journal of the Science of Food and Agriculture 28, 5358. MacRitchie, F., 1978. Differences in baking quality between wheat ours. Journal of Food Technology 13, 187194. MacRitchie, F., 1990. Chemistry at Interfaces. Academic Press, Inc., San Diego, CA, USA. MacRitchie, F., Gras, P.W., 1973. The role of our lipids in baking. Cereal Chemistry 50, 292302. Mills, E.N.C., Wilde, P.J., Salt, L.J., Skeggs, P., 2003. Bubble formation and stabilization in bread dough. Food and Bioproducts Processing 81, 189193. Paternotte, T.A., Orsel, R., Hamer, R.J., 1993. Interactions between our proteins and our lipids at the liquid/air interface. In: Gluten Proteins. Association of Cereal Research, Detmold, Germany, pp. 207217. Ponte Jr., J.G., DeStefanis, V.A., 1969. Note on separation and baking properties of polar and nonpolar wheat our lipids. Cereal Chemistry 46, 325329. Ross, A., MacRitchie, F., 1995. Interactions of wheat proteins, carbohydrates and lipids. In: Gaonkar, A.G. (Ed.), Ingredient Interactions Effects on Food Quality. Marcel Deker, New York, USA, pp. 321356. Salt, L.J., Wilde, P.J., Georget, D., Wellner, N., Skeggs, P.K., Mills, E.N.C., 2006. Composition and surface properties of dough liquor. Journal of Cereal Science 43, 284292.
glutenstarch matrix and that they act purely as surface active compounds at these levels of addition. 4. Conclusion This study for the rst time demonstrates that the lipids at their natural levels do not affect dough rheology and gas cell stabilization by the glutenstarch matrix. Nevertheless, their ability to modify loaf volume and crumb structure supports the dual lm hypothesis of Gan et al. (Gan et al., 1990, 1995). It suggests the presence of liquid lamellae, providing an independent mechanism of gas cell stabilization. The effects of different surface active components may be explained by the type of monolayer that they form. References
Baker, J.C., Parker, H.K., Mize, M.D., 1946. Supercentrifugates from dough. Cereal Chemistry 23, 539544. DeStefanis, V.A., Ponte Jr., J.G., 1976. Studies on breadmaking properties of wheatour nonpolar lipids. Cereal Chemistry 53, 636642. Dobraszczyk, B.J., Smewing, J., Albertini, M., Maesmans, G., Schoeld, J.D., 2003. Extensional rheology and stability of gas cell walls in bread doughs at elevated temperatures in relation to breadmaking performance. Cereal Chemistry 80, 218224. Gaines, G.L., 1966. Insoluble monolayers at liquidgas interfaces. In: Prigogine, I. (Ed.), Interscience Monographs on Physical Chemistry. John Wiley and Sons, Inc., New York, USA. Gan, Z., Angold, R.E., Williams, M.R., Ellis, P.R., Vaughan, J.G., Galliard, T., 1990. The microstructure of gas retention of bread dough. Journal of Cereal Science 12,1524. Gan, Z., Ellis, P.R., Schoeld, J.D., 1995. Gas cell stabilization and gas retention in wheat bread dough. Journal of Cereal Science 21, 215230.

Expression of globulin-2, a member of the cupin superfamily of proteins with similarity to known food allergens, is increased under high temperature regimens during wheat grain development
Susan B. Altenbach*, Charlene K. Tanaka, William J. Hurkman, William H. Vensel
USDA-ARS Western Regional Research Center, 800 Buchanan Street, Albany, CA 94710, USA
Article history: Received 27 February 2008 Received in revised form 28 June 2008 Accepted 8 July 2008 Keywords: Flour quality Gene expression Proteomics Quantitative RT-PCR
a b s t r a c t
Twenty-three expressed sequence tags (ESTs) from the US spring wheat Butte 86 were identied that encode proteins similar to a globulin-2 protein from maize embryos. The ESTs assembled into three contigs, two of which include the entire coding region for the mature protein. The encoded proteins contain two cupin domains and show signicant identities with 7S seed proteins from other species that are known or putative food allergens. Quantitative reverse-transcriptase polymerase chain reaction (qRTPCR) with primers specic for two of the sequences demonstrated that the globulin-2 genes are expressed late in grain development and that transcript levels increase when grain is produced under high temperature conditions. Transcripts were detected in both whole grain and endosperm, but levels were signicantly higher in whole grain and highest in embryo. In wheat our, at least 17 protein spots that differ in both size and pI were identied as globulin-2 by 2-DE/MS. Seven of the spots increased more than 2-fold in relative proportion when grain was produced under high temperature regimens. The data suggest that both transcriptional and post-translational mechanisms are involved in the response of globulin-2 to high temperatures. Published by Elsevier Ltd.
1. Introduction Wheat is one of the major crops grown throughout the world with a primary use in human nutrition. In 2005, wheat was grown on more than 200 million hectares and more than 600 million tons of grain were produced worldwide (http://faostat.fao.org/). A large variety of breads, noodles and baked goods are made from wheat our because of the unique viscoelastic properties that result when the our is mixed with water. These viscoelastic properties are determined largely by the gluten proteins, the major storage proteins that comprise 8085% of wheat our protein. The gluten proteins, consisting of gliadins and glutenin subunits, are characterized by unusually high levels of proline and glutamine and have very low solubilities in water or salt solutions. These proteins have been studied extensively because of their importance in our functionality. The remaining 1520% of the endosperm protein is
Abbreviations: 2-DE, two-dimensional gel electrophoresis; 2-DE/MS, twodimensional gel electrophoresis/mass spectrometry; DPA, days post-anthesis; ESTs, expressed sequence tags; MS/MS, tandem mass spectrometry; NPK, nitrogenphosphorouspotassium; qRT-PCR, quantitative reverse-transcriptase polymerase chain reaction; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis. * Corresponding author. Tel.: 1 510 559 5614; fax: 1 510 559 5818. E-mail address: altnbach@pw.usda.gov (S.B. Altenbach). 0733-5210/$ see front matter Published by Elsevier Ltd. doi:10.1016/j.jcs.2008.07.005
a heterogeneous mixture of proteins, most of which are soluble in water or dilute salt solutions. Within the salt-soluble protein fraction are a number of proteins related to the predominant storage proteins characterized in seeds from dicotyledonous plants. Some of these proteins have been shown to have sedimentation values similar to the 7S vicilins from pea and to share antigenic determinants with the pea vicilins (Robert et al., 1985). A 7S 3S globulin fraction characterized by Robert et al. (1985) contained major proteins of w75, 55, 3640, 24, 1720 and 10 kDa when analyzed by SDS-PAGE. Recently, a 2-DE/MS approach was used to identify more than 200 proteins contained in a salt-soluble fraction of wheat endosperm from the US spring wheat Butte 86 (Vensel et al., 2005). Seven spots in the 55 kDa region of the 2-D gels were determined to be most similar to a 7S vicilin-like protein characterized from maize embryos, referred to as globulin-2 (Wallace and Kriz, 1991). These proteins were found in wheat endosperm at 36 DPA but not at 10 DPA. In a subsequent proteomic analysis of endosperm from developing grain, the globulin-2 proteins were shown to accumulate during late stages of wheat grain development. Interestingly, these proteins also were among a group that increased in relative proportion when grain development occurred under a high temperature regimen of 37/28 C (day/night) rather than under a moderate temperature regimen of 24/17 C (Hurkman et al., 2009). Since these data suggest biological roles for the globulin-2 proteins
48
S.B. Altenbach et al. / Journal of Cereal Science 49 (2009) 4754
in the response of the grain to abiotic stress, we surveyed databases of wheat expressed sequence tags (ESTs) to identify the complement of globulin-2 genes from the US spring wheat Butte 86. Additionally, we examined the expression of two of the globulin-2 genes in developing grain and endosperm under different controlled environmental regimens using quantitative RT-PCR and evaluated the relative proportions of globulin-2 proteins in our produced from grain subjected to moderate and high temperature regimens. 2. Experimental
value was determined for each RNA sample and primer pair by the iCycler software. Mean normalized expression and standard errors were determined for each target gene relative to the 18S reference using the Q-Gene Core Module (available at http://www.genequantication.de/download.html#qgene). Amplication efciencies of each primer pair were taken into account for all calculations using equation #3 of Muller et al. (2002). Mean normalized expression was plotted as a function of chronological age for each experiment.
2.3. Protein extraction, 2-DE and MS/MS analyses 2.1. Growth of plants and tissue collection The US hard red spring wheat Triticum aestivum cv. Butte 86 was grown in a climate-controlled greenhouse as described by Altenbach et al. (2007). During grain development, plants were placed under either a moderate (24/17 C day/night) or a high (37/28 C day/night) temperature regimen. Levels of fertilizer (Plantex 202020 NPK) were adjusted at anthesis so that plants received 300 mg per day (1), 150 mg per day (0.5) or no (0) NPK during grain development. Developing grain and endosperm were collected at various intervals after anthesis under both temperature regimens. Embryo tissue was harvested from 30 DPA grain from plants grown under the 24/17 C regimen. Awns, glumes, stems and leaves were obtained during middle stages of grain development under the moderate regimen. Roots were harvested from seedlings grown on water-saturated sand at room temperature for 20 days. All tissues were frozen in liquid nitrogen and stored at 80 C until use. 2.2. RNA preparation and qRT-PCR Total RNA was isolated from all tissues as described previously for endosperm (Altenbach, 1998), treated with RQ1 RNAse-free DNAse (Promega, Madison, WI), and further puried by extraction with phenol:chloroform:isoamylalcohol (24:24:1) and ethanol precipitation. RNA was reverse-transcribed using the QuantiTect Reverse Transcription Kit (QIAGEN, Valencia, CA) according to the manufacturers directions. Amplication reactions were carried out in a volume of 25 ml containing cDNA, 0.3 mM of forward and reverse primers and SYBR Green Supermix (Biorad Laboratories, Hercules, CA) using a BioRad iCycler with an initial denaturation step of 95 C for 3 min, followed by 40 cycles of 95 C for 10 s and 55 C for 45 s. At the end of the PCR cycles, a melting curve was generated and analyzed by the iCycler software. Oligonucleotide primers for globulin-2 genes were based on expressed sequence tags (ESTs) from developing grain of Butte 86 available from NCBI. Primers were selected using Beacon Designer 4.0 (Premier Biosoft International, Palo Alto, CA) and synthesized by Operon (Huntsville, AL). Forward primer QF123 has the sequence 50 GGTCCAAATGGCTCCGC 30 and reverse primer QR123 has the sequence 50 CAAGAAGGACTGCGGGC 30 . Forward primer QF126 has the sequence 50 GGTCGTCATGCTCCTCAAC 30 and reverse primer QR126 has the sequence 50 GACGCTGAAGAAGGACTGTG 30 . Primers for an 18S rRNA used as a reference were reported in Altenbach et al. (2007). Amplication efciencies for each primer pair were calculated from standard curves generated in three independent experiments using the iCycler software. Each standard curve had a minimum of ve points and R values greater than 0.99. Amplication products were evaluated by gel electrophoresis on 4% Metaphor agarose gels (VWR International, Brisbane, CA). For quantication of transcripts, PCR reactions were carried out in triplicate using cDNA from the equivalent of 10 ng RNA from each time point. The 18S rRNA served as a reference RNA and was amplied in parallel with target genes in all experiments. The Ct Endosperm was collected from developing grain and a KClsoluble/methanol-insoluble protein fraction was isolated and quantied as described previously (Hurkman and Tanaka, 2004; Vensel et al., 2005). Equal amounts of protein (18 mg) from each time point were separated by 2-DE in triplicate. Gels were digitized with a calibrated scanner (UMAX PowerLook III, Dallas, TX) at 300 dpi using the same setting for all gels and analyzed using a computerized image analysis system (Progenesis PG240 version 2006, Non-Linear Dynamics Limited, Newcastle upon Tyne, UK). Normalized volumes (individual spot volume/total spot volume 100) were determined for each spot, averaged over the three gels, and plotted against the chronological age of the grain in DPA. Grain produced under different temperature regimens was tempered to 15% moisture and milled to our by the USDA-ARS Wheat Quality Laboratory (Manhattan, KS) using a Brabender Quadrumat Jr. mill (Hackensack, NJ). Methods for extraction and analysis of proteins from our were the same as those used for endosperm. For identication of proteins by tandem mass spectrometry (MS/MS), spots were excised from a representative gel and reduction, alkylation, reagent removal, and tryptic digestion were carried out automatically by a DigestPro xyz robot (INTAVIS Bioanalytical Instruments AG, Bergisch Gladbach, Germany). The DigestPro collection tray containing the tryptic peptides was placed in an autosampler that was interfaced with a QSTAR pulsar I hybrid quadrupole-TOF instrument (Applied Biosystems/MDX Sciex, Toronto, Canada) congured with an ESI source. Acquisition of tandem mass spectrometry data was as follows. From an initial survey scan of mass range m/z 4002000, the most abundant doubly or triply charged ion above a threshold of 20 counts was selected for fragmentation. Collision induced dissociation of the mass-selected ion was carried out using UHP nitrogen. Following the 3S MS/MS fragmentation period, the MS survey scan was repeated until another MS/MS period was triggered. Wiff data les were created for each sample by the Analyst QS version 1.1 software, converted to MGF text les using Mascot Daemon (http:// www/matrixscience.com/) and submitted in batch mode to a locally installed copy of X!Tandem (Craig and Beavis, 2004) using a script provided by Jayson Faulkner (University of Michigan). X!Tandem was congured to search les containing sequences from the NCBI non-redundant collection: nr-Arabidopsis-thaliana.fasta, nr-other-Viridiplantae.fasta, nr-Oryza-sativa.fasta, in addition to entries from all proteins in the HarvEST:Wheat version 1.04 database (http://harvest.ucr.edu/HWheat104.exe), NCBI Triticum aestivum: UniGene Build #37, and wEST Database (http://wheat.pw. usda.gov/wEST). The results were visualized using a locally installed copy of the Global Proteome Machine (GPM) (http:// thegpm.org/). Trypsin was selected as the cleavage enzyme. The results were searched with a fragment ion mass tolerance of 1.00 Da and a parent ion tolerance of 0.40 Da. Oxidation of methionine was specied as a variable modication (see also: Vensel et al., 2002, 2005). Identications accepted as valid had log e values less than 4.0.
49
3. Results 3.1. Accumulation of globulin-2 in developing endosperm under different temperature regimens A number of protein spots were identied as globulin-2 in proteomic maps of KCl-soluble/MeOH-insoluble proteins from developing endosperm from Butte 86 (Vensel et al., 2005). Six spots (referred to as 147, 155, 559, 561, 851, and 852 in Vensel et al., 2005) ranged from about 54.8 to 56.1 kDa by 2-DE and had pIs between 6.2 and 7.0. Another spot (631 in Vensel et al., 2005) was about 15 kDa and had a pI of about 5.7. Accumulation proles for these seven globulin-2 proteins during grain development under moderate and high temperature regimens were summarized as part of a large-scale proteomic analysis by Hurkman et al. (2009) and are shown in more detail in Fig. 1. Under moderate temperatures (24/17 C), these proteins accumulated very late in grain development, between 32 and 34 DPA, and each protein spot
encompassed less than 0.2% of the normalized volumes of all spots in the KCl-soluble/MeOH-insoluble fraction. When a high temperature regimen of 37/28 C was applied from 10 DPA until maturity, the proteins accumulated earlier, between 22 and 24 DPA. In addition, maximum normalized spot volumes were 1.9- to 4.1-fold higher than in grain produced under the 24/17 C regimen (Fig. 1 and Hurkman et al., 2009). 3.2. Identication of ESTs encoding globulin-2 from Butte 86 A collection of contigs assembled from more than 116,000 wheat ESTs by Chao et al. (2006) was surveyed to identify sequences that encode the wheat globulin-2 proteins. The EST collection includes 3639 ESTs from Butte 86 developing grain. Information about both contigs and ESTs is available at http://wheat.pw.usda.gov/westsql/. Three contigs, NSFT03P2_Contig18428, NSFT03P2_Contig17295 and NSFT03P2_Contig17366, encode proteins similar to a globulin-2 characterized previously in maize (Wallace and Kriz, 1991). Amino
Fig. 1. Accumulation of globulin-2 proteins in endosperm from grain produced under moderate and high temperatures. In each panel the solid lines denote endosperm produced under the 24/17 C regimen and the dashed lines denote endosperm produced under the 37/28 C regimen. The high temperature regimen was imposed from 10 DPA to maturity and plants were supplied with 0.5 NPK. Bars indicate the standard error for each time point among triplicate gels. Accumulation data for spots 155, 559, 851, 147, 631, 852 and 561 from Vensel et al. (2005) are presented in panels A, B, C, D, E, F and G, respectively.
50
acid sequences of the proteins encoded by the contigs are shown in Fig. 2. Contig 18428 includes nine ESTs from Butte 86 (BQ806323, BQ807065, BQ806087, BQ805228, BQ806420, BQ805507, BQ806904, BQ804552, BQ804667) that represent the entire coding region. Contig 17295 includes four ESTs from Butte 86 (BQ805639, BQ839104, BQ805390, BQ807159) that cover the 50 portion of the coding region and one EST (BQ805520) that represents the 30 end. Nine ESTs from Butte 86 were included in contig 17366 (BQ839060, BQ805920, BQ807178, BQ838786, BQ839055, BQ838678, BQ807089, BQ806653, BQ805350) and all nine represent the 50 half of the coding region. It is likely that this contig is missing the 30 end of the coding region. When evaluated with the Signal P algorithm (http://www.cbs. dtu.dk/services/SignalP/), proteins encoded by the three contigs are predicted to contain signal peptide cleavage sites between the alanine and the serine residues at positions 26 and 27 in contigs 17295 and 17366 and positions 20 and 21 in contig 18428 (Bendtsen et al., 2004), suggesting that these proteins enter the secretory pathway (Fig. 2). Contigs 18428 and 17295 encode proteins with predicted molecular weights of 53,495 and 53,773, respectively, excluding the signal peptide. Both proteins contain abundant arginine, glycine and glutamine residues. An InterProScan (http:// www.ebi.ac.uk/InterProScan/) showed that each protein contains two cupin domains (Fig. 2) that are able to form barrel-like structures typical of the cupin superfamily of proteins. Since many plant proteins within the cupin superfamily are food allergens (Mills et al., 2002), the protein encoded by contig 17295 was used to search the Food Allergy Research and Resource Program (FARRP) Protein AllergenOnline Database version 8.0 (http://www. allergenonline.com/databasefasta.asp). A sliding 80-mer search revealed matches of >35% identity between this globulin-2 and 18 7S proteins that are known or putative allergens (Table 1). These include proteins from sesame, various tree nuts (English walnut, black walnut, hazelnut, cashew) and legumes (lentil, pea, soybean, peanut). Current Codex Alimentarius guidelines indicate that proteins with identities to known allergens that are >35% over 80 amino acid regions may be allergenic (Goodman, 2006).
3.3. Selection of primers for qRT-PCR and analysis of gene expression ESTs BQ806323, BQ805639 and BQ839060 from cv Butte 86 represent the rst half of the coding sequences of contigs 18428, 17295 and 17366, respectively, and were used for the design of gene-specic primers. Primer QF123 was a perfect match with BQ806323, but had only 13/17 nucleotides identical to BQ839060 and BQ805639 (Fig. 3). Primer QR123 was a reverse complement of BQ806323, but had only 13/17 nucleotides that complemented either BQ839060 or BQ805639. QF123/QR123 amplied a 110 bp fragment that was veried to correspond to BQ839060 by digestion with AlwI, HaeII, and HpaII. The average amplication efciency of QF123/QR123 was 96.7%. Primer QF126 was a perfect match with both BQ805639 and BQ839060 but had only 16/19 nucleotides in common with BQ806323 (Fig. 3). Primer QR126 was a reverse complement of BQ805639, but had only 18 and 17 of 20 residues that complemented BQ806323 and BQ839060, respectively. QF126/ QR126 amplied a 100 bp fragment that was veried to correspond to BQ805639 by digestion with HpaII, MboII, AciI, AlwI and HaeII. The average amplication efciency of QF126/QR126 was 93.7%. Transcript accumulation for the two globulin-2 genes was assessed in developing grain produced under either moderate or high temperature regimens from anthesis to maturity (Fig. 4). In the absence of post-anthesis NPK (Fig. 4A, B), transcripts corresponding to both genes were rst detected at 11 DPA under moderate temperatures and reached maximum levels at 26 DPA. Under the high temperature regimen, transcripts began to accumulate earlier in grain development, by 6 DPA, and reached maximum levels at 22 DPA. Changes in the timing of globulin-2 transcript accumulation were consistent with changes in the timing of developmental events due to high temperature (Altenbach et al., 2003) and with changes in the timing of expression of other genes (Altenbach and Kothari, 2004; Altenbach et al., 2007, 2008). Maximum levels of transcripts corresponding to BQ806323 and BQ805639 were 4.1and 4.9-fold higher, respectively, under high temperatures than under moderate temperatures. In a second growth experiment,
Fig. 2. Comparison of protein sequences encoded by NSFT03P2_Contig18428, NSFT03P2_Contig17295 and NSFT03P2_Contig17366. Identical amino acid residues in each protein are enclosed in boxes. Brackets denote portions of the coding regions in each contig that correspond to ESTs from Butte 86. The arrow indicates putative signal peptide cleavage sites. Cupin domains in each protein are shown in bold type and putative N-linked glycosylation sites are indicated by asterisks.
S.B. Altenbach et al. / Journal of Cereal Science 49 (2009) 4754 Table 1 Comparison of globulin-2 encoded by NSFT03P2_Contig17295 with known and putative allergens in the FARRP Protein AllergenOnline Database Version 8.0 using a sliding 80-mer search GenBank accession # 13183177 6580762 31321944 19338630 21666498 21914823 29539109 46560474 42414627 42414629 1168390 46560472 1168391 46560476 29539111 256427 18536 169929 Source Sesame indicum (sesame) Juglans regia (English walnut) Juglans nigra (black walnut) Corylus avellana (hazelnut) Anacardium occidental (cashew) Anacardium occidental (cashew) Lens culinaris (lentil) Arachis hypogaea (peanut) Pisum sativum (pea) Pisum sativum (pea) Arachis hypogaea (peanut) Arachis hypogaea (peanut) Arachis hypogaea (peanut) Arachis hypogaea (peanut) Lens culinaris (lentil) Glycine max (soybean) Glycine max (soybean) Glycine max (soybean) Best % identity 51.90 48.10 47.50 47.50 43.80 43.80 42.51 41.29 41.24 41.24 40.70 40.70 40.70 40.70 40.00 38.79 36.29 35.03 # Hits > 35%a 285 269 251 204 173 179 112 62 97 105 75 58 79 79 123 45 11 2
51
Proteins with >35% identity over 80 amino acids to wheat globulin-2 are shown. a of 426 80-mers.
and BQ805639, respectively, than at the 34 DPA time point under moderate temperatures. In the experiments shown in Fig. 4, globulin-2 transcript levels relative to the 18S reference RNA were higher in whole grain than in endosperm. Fig. 5 shows a direct comparison of globulin-2 transcript levels in whole grain at 30 DPA and in embryo and endosperm dissected from 30 DPA grain produced under the moderate temperature regimen. In this analysis, transcripts corresponding to BQ806323 and BQ805639 were 10.1- and 7.8-fold higher, respectively, in whole grain than in endosperm. BQ806323 and BQ805639 transcripts were most abundant in embryo, about 2.5- and 2.6-fold higher, respectively, in embryo than in whole grain. Consistent with these ndings is the fact that a signicant proportion of the ESTs that comprise the globulin-2 contigs were derived from an embryo cDNA library, including 9/22 for contig 18428, 4/11 for contig 17295 and 2/11 for contig 17366. The collection of ESTs assembled into contigs by Chao et al. (2006) included 4147 sequences from endosperm and 1916 sequences from embryo in addition to the 3639 from Butte 86 whole grain. RNA prepared from awns, glumes, stems, leaves and roots also was analyzed in the same experiment. Transcripts corresponding to BQ805639 were detected at very low levels in roots (Fig. 5B), but were not detectable in awns, glumes, stems or leaves. BQ806323 was not expressed in any of the nongrain samples (data not shown).
plants were supplied with post-anthesis NPK (Fig. 4C, D). In this experiment, transcripts for BQ806323 were rst detected by 12 DPA and reached maximum levels at 34 DPA under moderate temperatures (Fig. 4C) while transcripts for BQ805639 were detectable at low levels at 7 DPA and reached maximum levels at 34 DPA (Fig. 4D). Under the high temperature regimen, transcripts for both genes were rst detected at 7 DPA and reached maximum levels at 18 DPA, the last time point sampled. Maximum levels of BQ806323 and BQ805639 transcripts were 4.1- and 2.1-fold higher, respectively, under the high temperature regimen than under moderate temperatures. The accumulation of globulin-2 transcripts also was assessed in developing endosperm (Fig. 4E, F). Transcripts for both BQ806323 and BQ805639 were detected at low levels at both 7 and 14 DPA under moderate temperatures and reached maximum levels at 34 DPA. When high temperatures were applied from anthesis to maturity, transcripts for both genes were detected at the rst time point sampled at 5 DPA and increased to maximum levels by 20 DPA. Transcript levels at the 20 DPA time point under the high temperature regimen were 3.9- and 5.3-fold higher for BQ806323
3.4. Globulin-2 proteins in our produced under moderate and high temperatures A proteomic analysis was undertaken to determine the levels of globulin-2 proteins in white our milled from grain produced under moderate and high temperatures with and without postanthesis NPK. This analysis revealed a complex collection of globulin-2 proteins in the KCl-soluble/MeOH-insoluble fraction (Fig. 6AC). Seventeen spots in the 5255 kDa region of the 2-D gels were identied as globulin-2. Spots 219, 209, 199, 200, 193, 190 and 258 in Fig. 6C correspond to spots 147, 155, 559, 561, 851, 852 and 175 identied in Vensel et al. (2005). MS/MS data used to identify the remaining spots (192, 202, 205, 221, 222, 228, 236, 239, 320, and 210) as globulin-2 proteins are summarized in Table S1. One of the spots (210) contained more than one protein and was excluded from further analysis. Collectively, the 16 remaining spots comprise about 2.9 and 3.1% of the total protein in the fraction when grain was produced under a moderate temperature regimen with or without post-anthesis NPK, respectively. When grain was produced under a high temperature regimen from anthesis to maturity, the 16 spots comprised a larger percentage of the total protein in the fraction, 5.2% when plants were supplied with NPK and 5.0% when plants did not receive post-anthesis NPK. Of the 16 spots, nine showed statistically signicant increases with high temperatures (P < 0.05) under at least one of the fertilizer regimens. Seven spots increased more than 2-fold in response to high temperature under both fertilizer regimens (192, 193, 199, 202, 205, 209 and 320) (Fig. 6C, D). The largest change with high temperature was a 8.3fold increase detected for spot 199 when grain was produced without NPK. When grain was produced with NPK, the largest response was a 4.8-fold increase for spot 320. Four spots (200, 221, 228 and 239) showed small decreases with high temperatures. The other three spots identied as globulin-2 proteins encompassed similar proportions of the KCl-soluble/MeOH-insoluble our fraction under all four environmental regimens (222, 236 and 258). Five of the seven spots that increased more than 2-fold (193, 199, 202, 209 and 320) also increased in our produced under high temperatures in a second growth experiment in which plants were supplied with 0.5 NPK (data not shown).
Fig. 3. Comparison of the nucleotide sequences of BQ806323, BQ839060 and BQ805639 in regions used for gene-specic primers. QF123 is identical to nucleotides 481497 of BQ806323 and QR123 is the reverse complement of nucleotides 573590 of BQ806323. QF126 is identical to nucleotides 532550 of BQ805639 and QR126 is the reverse complement of nucleotides 612631 of BQ805639. Nucleotides that differ in other ESTs are underlined.
52
Fig. 4. Accumulation of transcripts corresponding to BQ806323 (A, C, E) and BQ805639 (B, D, F) in developing grain (AD) or endosperm (E, F) produced under a 24/17 C regimen (solid lines) or a 37/28 C regimen (dashed lines). Plants in A and B did not receive post-anthesis NPK, while plants in C and D received 1 NPK and plants in E and F received 0.5 NPK. Bars indicated the standard error among triplicate reactions.
4. Discussion In dicots, the 7S globulins are embryo proteins that serve a storage function and are the primary source of amino acids for the developing seedling. In wheat, the 7S globulins make relatively small contributions to the storage reserves of the grain and the primary storage proteins are the gliadins and glutenin subunits. While Halford and Shewry (2007) suggested that globulin proteins
in cereals have no biological role other than storage, data presented here suggests that some of the wheat globulin-2 proteins may be involved in the response of the developing grain to high temperature stress. Transcript proles demonstrate that two closely-related globulin-2 genes are expressed in whole grain and endosperm late in grain development and that levels of transcripts increase in response to high temperatures. Globulin-2 proteins also encompass a greater proportion of the total KCl-soluble/MeOH-insoluble
Fig. 5. Accumulation of transcripts corresponding to BQ806323 (A) and BQ805639 (B) in embryo, endosperm and whole grain, all at 30 DPA, and in root. Bars indicate the standard error among triplicate reactions.
53
Fig. 6. 2-DE analysis of proteins from our produced under a 24/17 C (A) and a 37/28 C (B) regimen with 1 NPK. Boxes indicate regions of each gel that contain spots identied by MS/MS as globulin-2. The boxed region of the gel from panel B is enlarged in panel C and spots identied as globulin-2 are numbered. Numbers that are underlined refer to spots that increase in relative proportion under high temperature conditions. Panel D compares the average normalized volumes for each globulin-2 spot from plants subjected to the 24/ 17 C and 37/28 C regimens, with and without post-anthesis NPK. The solid and open bars denote the 24/17 C regimens with or without NPK, respectively. The hatched and stippled bars denote the 37/28 C regimen, with or without NPK, respectively. Asterisks denote changes with high temperature under each NPK regimen that are statistically signicant (P < 0.05).
protein in endosperm and our from grain produced under high temperature conditions. These ndings are consistent with previous work on the homologous globulin-2 protein from maize. In an early study, Kriz and Wallace (1991) reported maize variants lacking globulin-2 and suggested that this protein was not essential for seed development, maturation or germination. More recent studies using proteomic approaches have shown that globulin-2 proteins respond to stress in maize. Kollipara et al. (2002) found that globulin-2 was among the proteins up-regulated in desiccation-tolerant maize hybrids and Labra et al. (2006) noted that a number of proteins identied as globulin-2 were up-regulated in maize plantlets germinated in the presence of potassium dichromate. Analysis of wheat ESTs indicates that the wheat globulin-2 proteins are encoded by a simple gene family consisting of three members. Yet at least 17 protein spots have been identied as globulin-2 in wheat our using 2-DE/MS, suggesting that the proteins undergo post-translational modications that result in changes in size and pI. It is likely that the wheat globulin-2 proteins, like 7S seed storage proteins from legumes, are glycosylated. Robert et al. (1985) noted charge heterogeneity among proteins isolated from a 7S 3S fraction of globulins from wheat grain in the 55 kDa region of 2-D gels and observed that 55 kDa proteins also were bound to a concanavalin A sepharose column. Additionally, proteins encoded by contigs 18428, 17295 and 17366 contain putative sites for N-linked glycosylation, dened as NXT, where X can be any amino acid except P (Fig. 2). Using afnity blotting, Kriz (1989) also noted that the homologous globulin-2 protein from maize was able
to bind Con A. Some 7S globulins also undergo proteolytic pro cessing (reviewed by Muntz, 1996). It is interesting that not all of the globulin-2 protein spots identied by 2-DE increased in response to high temperatures. In fact, decreases in the levels of several globulin-2 proteins were detected in our. Thus, both transcriptional and post-translational mechanisms may be involved in the response of globulin-2 genes to high temperatures in the developing grain. Transcripts for globulin-2 were present at lower levels in the endosperm than in whole grain. Halford and Shewry (2007) noted that 7S globulins from cereals are expressed predominantly in the embryo and in the aleurone layer of the endosperm. Our analyses did not distinguish the aleurone tissue, which is a single cell layer in wheat, from the starchy endosperm tissue. Nonetheless, transcript levels clearly increased in both the whole grain and the endosperm samples. Globulin-2 proteins were present in endosperm tissue extruded from grain from which the embryo had been removed. In addition, globulin-2 proteins were present in milled white our that generally contains minimal amounts of germ and bran. Increases in the relative proportion of globulin-2 proteins in our produced under high temperatures may have important implications given the potential allergenicity of these proteins. The KCl-soluble/MeOH-insoluble proteins account for about 11% of the total our protein (Hurkman and Tanaka, 2004) and globulin-2 proteins comprise as much as 6% of this fraction under high temperature conditions. Wheat is among the top eight foods responsible for human food allergies. Allergies to wheat are characterized by a broad spectrum of symptoms including urticaria,
54
S.B. Altenbach et al. / Journal of Cereal Science 49 (2009) 4754 Altenbach, S.B., Kothari, K.M., Tanaka, C.K., Hurkman, W.J., 2007. Genes encoding the PR-4 protein wheatwin are developmentally regulated in wheat grains and respond to high temperatures during grainll. Plant Science 173, 135143. Altenbach, S.B., Kothari, K.M., Tanaka, C.K., Hurkman, W.J., 2008. Expression of 9kDa non-specic lipid transfer protein genes in developing wheat grains is enhanced by high temperatures but not by post-anthesis fertilizer. Journal of Cereal Science 47, 201213. Battais, F., Courcoux, P., Popineau, Y., Kanny, G., Moneret-Vautrin, D.A., DeneryPapini, S., 2005. Food allergy to wheat: differences in immunoglobin Ebinding proteins as a function of age or symptoms. Journal of Cereal Science 42, 109117. Bendtsen, J.D., Nielsen, H., von Heijne, G., Brunak, S., 2004. Improved prediction of signal peptides: SignalP 3.0. Journal of Molecular Biology 340, 783795. Chao, S., Lazo, G.R., You, F., Crossman, C.C., Hummel, D.D., Lui, N., LaudenciaChingcuanco, D., Anderson, J.A., Close, T.J., Dubcovsky, J., Gill, B.S., Gill, K.S., Gustafson, J.P., Kianian, S.F., Lapitan, N.L.V., Nguyen, H.T., Sorrells, M.E., McGuire, P.E., Qualset, C.O., Anderson, O.D., 2006. Use of a large-scale Triticeae expressed sequence tag resource to reveal gene expression proles in hexaploid wheat (Triticum aestivum L.). Genome 49, 531544. Craig, R., Beavis, R.C., 2004. TANDEM: matching proteins with mass spectra. Bioinformatics 20, 14661467. Goodman, R.E., 2006. Practical and predictive bioinformatics methods for the identication of potentially cross-reactive protein matches. Molecular Nutrition and Food Research 50, 655660. Halford, N.G., Shewry, P.R., 2007. The structure and expression of cereal storage protein genes. Plant Cell Monographs 8, 195218. Hurkman, W.J., Tanaka, C.K., 2004. Improved methods for separation of wheat endosperm proteins and analysis by two-dimensional gel electrophoresis. Journal of Cereal Science 40, 295299. Hurkman, W.J., Vensel, W.H., Tanaka, C.K., Whitehand, L., Altenbach, S.B., 2009. Effect of high temperature on albumin and globulin accumulation in the endosperm proteome of the developing wheat grain. Journal of Cereal Science 49, 1223. Kollipara, K.P., Saab, I.N., Wych, R.D., Lauer, M.J., Singletary, G.W., 2002. Expression proling of reciprocal maize hybrids divergent for cold germination and desiccation tolerance. Plant Physiology 129, 974992. Kriz, A.L., 1989. Characterization of embryo globulins encoded by the maize Glb genes. Biochemical Genetics 27, 239251. Kriz, A.L., Wallace, N.H., 1991. Characterization of the maize globulin-2 gene and analysis of two null alleles. Biochemical Genetics 29, 241254. Labra, M., Gianazza, E., Waitt, R., Eberini, I., Sozzi, A., Regondi, S., Grassi, F., Agradi, E., 2006. Zea mays L. protein changes in response to potassium dichromate treatments. Chemosphere 62, 12341244. Mills, E.N.C., Jenkins, J., Marigheto, N., Belton, P.S., Gunning, A.P., Morris, V.J., 2002. Allergens of the cupin superfamily. Biochemical Society Transactions 30, 925929. Muntz, K., 1996. Proteases and proteolytic cleavage of storage proteins in developing and germinating dicotyledonous seed. Journal of Experimental Botany 47, 605622. Muller, P.Y., Janovjak, H., Miserez, A.R., Dobbie, Z., 2002. Processing of gene expression data generated by quantitative real-time RT-PCR. BioTechnology 32, 13721379. Robert, L.S., Adeli, K., Altosaar, I., 1985. Homology among 3S and 7S globulins from cereals and pea. Plant Physiology 78, 812816. Salt, L.J., Robertson, J.A., Jenkins, J.A., Mulholland, F., Mills, E.N.C., 2005. The identication of foam-forming soluble proteins from wheat (Triticum aestivum) dough. Proteomics 5, 16121623. Vensel, W.H., Harden, L., Tanaka, C.K., Hurkman, W.J., Haddon, W.F., 2002. Identication of wheat endosperm proteins by MALDI mass spectrometry and LC-MS/ MS. Journal of Biomolecular Techniques 13, 95100. Vensel, W.H., Tanaka, C.K., Cai, N., Wong, J.H., Buchanan, B.B., Hurkman, W.J., 2005. Developmental changes in the metabolic protein proles of wheat endosperm. Proteomics 5, 15941611. Wallace, N.H., Kriz, A.L., 1991. Nucleotide sequence of a cDNA clone corresponding to the maize globulin-2 gene. Plant Physiology 95, 973975.
gastrointestinal distress, asthma and anaphylaxis and are due to sensitivities to a variety of proteins in both the gluten protein and albumin/globulin fractions (Battais et al., 2005). Since the wheat globulin-2 shows signicant identity to other proteins that are known or putative allergens, further studies on the allergenicity of these proteins are warranted. IgE binding studies using sera from patients with characterized wheat allergies would be key to determining whether the globulin-2 proteins should be classied as food allergens. An additional question is whether changes in the relative proportions of the globulin-2 proteins might alter the functional properties of the our. It is generally accepted that the gluten proteins contribute elasticity and extensibility properties essential for the formation of wheat our doughs. Interactions between the gluten protein matrix and other our components also may affect rheological properties but have been little studied. Recently, the globulin-2 proteins were identied in a proteomic study of a dough liquor fraction prepared from wheat our (Salt et al., 2005). It is interesting that the levels of the globulin-2 proteins in the dough liquor increased when salt or salt plus ascorbate was included in the dough mixture. Salt is frequently included in commercial processes to improve breadmaking properties and has been shown to inuence water absorption, dough development time, dough stability and loaf volume. It is not known whether the globulin-2 proteins are associated with the gluten matrix in the absence of salt. Further study is required to determine whether these proteins play roles in our functionality. Acknowledgements The authors thank Drs. Frances DuPont and Ann Blechl for critical review of the manuscript. Mention of a specic product does not constitute an endorsement and does not imply a recommendation over other suitable products. Appendix. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jcs.2008.07.005. References
Altenbach, S.B., 1998. Quantication of individual low-molecular-weight glutenin subunit transcripts in developing wheat grains by competitive RT-PCR. Theoretical and Applied Genetics 97, 413421. Altenbach, S.B., DuPont, F.M., Kothari, K.M., Chan, R., Johnson, E.L., Lieu, D., 2003. Temperature, water and fertilizer inuence the timing of key events during grain development in a US spring wheat. Journal of Cereal Science 37, 920. Altenbach, S.B., Kothari, K.M., 2004. Transcript proles of genes expressed in endosperm tissue are altered by high temperature during wheat grain development. Journal of Cereal Science 40, 115126.

Biochemical markers: Efcient tools for the assessment of wheat grain tissue proportions in milling fractions
Youna Hemery a, Valerie Lullien-Pellerin a, Xavier Rouau a, Joel Abecassis a, Marie-Franoise Samson a, Per man b, Walter von Reding c, Cacilia Spoerndli c, Cecile Barron a, *
a b c
INRA, UMR 1208 Agropolymers Engineering and Emerging Technologies, INRA-CIRAD-UMII-Supagro, F-34000 Montpellier, France Department of Food Science, Swedish University of Agriculture Science (SLU), Box 7051, S-750 07 Uppsala, Sweden Buhler AG, CH-9240 Uzwil, Switzerland
Article history: Received 10 April 2008 Received in revised form 3 July 2008 Accepted 8 July 2008 Keywords: Wheat Grain Bran Aleurone Outer layers Fractionation Milling Processes
a b s t r a c t
To produce safe and healthy whole wheat food products, various grain or bran dry fractionation processes have been developed recently. In order to control the quality of the products and to adapt these processes, it is important to be able to monitor the grain tissue proportions in the different milling fractions produced. Accordingly, a quantitative method based on biochemical markers has been developed for the assessment of grain tissue proportions in grain fractions. Grain tissues that were quantied were the outer pericarp, an intermediate layer (including the outer pericarp, the testa and the hyaline layer), the aleurone cell walls, the aleurone cell contents, the endosperm and the germ, for two grain cultivars (Tiger and Crousty). Grain tissues were dissected by hand and analysed. Biochemical markers chosen were ferulic acid trimer, alkylresorcinols, para-coumaric acid, phytic acid, starch and wheat germ agglutinin, for outer pericarp, intermediate layer, aleurone cell walls, aleurone cell contents, endosperm and germ respectively. The results of tissue quantication by hand dissection and by calculation were compared and the sensitivity of the method was regarded as good (mean relative errors of 4% and 8% for Crousty and Tiger outer layers respectively). The impact of the analytical variability (maximum 13% relative error on coarse bran) was also regarded as acceptable. Wheat germ agglutinin seems to be a promising marker of wheat germ: even if the quantication method was not able to quantify the germ proportions in milling fractions, it was able to classify these fractions according to their germ content. The efciency of this method was tested, by assessing the grain tissue proportions of fractions exhibiting very different compositions such as our, bran and aleurone-rich fractions obtained from three different grain or bran dry fractionation processes (conventional milling, debranning process, production of aleurone-rich fractions from coarse bran). By calculation of the composition of the different products generated, it was possible to study the distribution of the different tissues among fractions resulting from the different fractionation processes. This quantitative method is thus a useful tool for the monitoring and improvement of processes, and allows the effects of these processes to be understood and their adaption to reach the objectives. 2008 Elsevier Ltd. All rights reserved.
1. Introduction Numerous epidemiological studies have demonstrated the health benets of consuming more whole-grain foods (Jacobs and Steffen, 2003; Liu, 2007). However, all the wheat grain parts are not health-promoting, for example the outermost parts have been
Abbreviations: ARs, alkylresorcinols; Cr, crousty cultivar; FAt, 4-O-80 , 50 -500 dehydrotriferulic acid (ferulic acid trimer); P2O5, phosphorus pentaoxyde; p-CA, para-coumaric acid; Tg, Tiger cultivar; WGA, wheat germ agglutinin. * Corresponding author. Tel.: 33 4 99 61 31 04; fax: 33 4 99 61 30 76. E-mail address: barron@supagro.inra.fr (C. Barron). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.07.006
shown to concentrate the majority of the grain contaminants, like microorganisms, mycotoxins, pesticide residues and heavy metals (Aureli and DEgidio, 2007; Fleurat-Lessard et al., 2007; Laca et al., 2006). On the other hand, the wheat aleurone layer has been shown to have great nutritional interest, and to concentrate most of the minerals and vitamins of the wheat grain (Pomeranz, 1988). Buri et al. (2004) reported that wheat aleurone layer contains interesting proportions of proteins, b-glucans, phenolic compounds and other phytochemicals (lignans, sterols). Zhou et al. (2004) pointed out that antioxidant, including phenolic acids, are concentrated in the aleurone layer of wheat bran, and Mateo Anson et al. (2008) showed that the higher the proportion of aleurone material in wheat fractions, the higher the antioxidant capacity observed for
56
Y. Hemery et al. / Journal of Cereal Science 49 (2009) 5564
these fractions. Amrein et al. (2003) found that aleurone-rich fractions exhibited better in vitro digestibility and colonic fermentability than wheat bran, and Bach Knudsen et al. (1995) observed that the digestibility of minerals, protein and non-starch polysaccharides is much higher in bran fractions rich in aleurone than in fractions rich in pericarp and testa. These results suggest that it could be interesting to produce aleurone-rich fractions for use as food ingredients. As a consequence, new processes are developed in order to exploit all the nutritional benets of whole grain and to produce new wheat foods and wheat-based ingredients with enhanced nutritional quality (Hemery et al., 2007). For example, depending on the desired product, a process can aim at discarding the pericarp to obtain whole grains containing less contaminants, while other processes may be developed in order to produce bran fractions highly concentrated in aleurone material (Buri et al., 2004). It is often difcult to exactly monitor the distribution of the different grain tissues among fractions during processing, as no simple method exists to quantify the respective proportions of these tissues in fractions. However, the monitoring of tissue proportions in the different fractions is essential, as it allows to control the quality of the products and to consequently adapt the processes. Therefore quantitative tools are needed. Different compounds can be measured to evaluate bran contamination in wheat ours and fractions. Ash content and measurement of our colour are widely used in the milling industry as indicators of our purity. The amino acid composition of the various tissues was also studied but did not allow the quantication of the grain tissues in milling fractions (Jensen and Martens, 1983). Some authors (Lempereur et al., 1998; Pussayanawin et al., 1988) suggested using the concentration of ferulic acid to quantify bran in ours and semolinas, and alkylresorcinols have more recently been shown to be good markers of wheat bran content in foods (Chen et al., 2004). Such analyses may be useful to evaluate the total outer layer content but they do not allow to distinguish between the different outer tissues. Another way to evaluate the different grain tissues in wheat fractions consists of the use of specic uorescence properties of the outer layers. Indeed, the aleurone cell walls display blue uorescence under UV-light, due to the presence of ferulic acid, whereas the pericarp shows green uorescence under blue light (Fulcher et al., 1972; Jensen et al., 1982; Symons and Dexter, 1996). Based on these uorescence properties, commercial equipment has been developed to determine the amount of aleurone in ours. Multispectral uorescence image analysis of grain sections coupled with classication techniques have also been developed to more precisely quantify the proportions of the different parts of the grain (Baldwin et al., 1997; Courcoux et al., 2002), but have not yet been applied to powdery samples. These imaging methods can be good tools for on-line use, but their main disadvantage would be their lack of specicity, as they do not allow to quantify other tissues than aleurone and pericarp. Moreover, all these methods allow determination of bran proportions in ours during milling, but they may not be adaptable to other fractionation systems (such as progressive abrasion or bran fractionation). Peyron et al. (2002) and Antoine et al. (2004) carried out biochemical analyses of isolated wheat grain tissues and used the differences in chemical composition between these tissues to assess the histological composition of technological fractions. They used compounds such as phenolic acids, phytic acid, and starch as biochemical markers. Indeed, these compounds were shown to be either present exclusively in one part of the grain (starch in starchy endosperm), or present in greater amounts in one particular tissue (phytic acid in aleurone cell contents and some phenolic acids in the cell walls). These biochemical markers were used to determine the amounts of aleurone layer and pericarp in ours and other milling fractions (Peyron et al., 2002), and in the samples obtained from a wheat bran fractionation process (Antoine et al., 2004).
Antoine et al. (2004) used different markers for aleurone cell walls and aleurone cell contents to assess the histological composition of bran fractions and to evaluate the dissociation and the accessibility of aleurone cellular components. This method was reported to provide accurate quantication of the histological composition of samples and to be versatile, as it can be rened and adapted depending on the type of sample analysed, from either bran or whole-grain fractionation. However, it did not allow the quantication of testa (this tissue was deduced by subtraction and thus was perhaps overestimated), and it neither allowed the detection of the germ. Having a marker for wheat germ would nevertheless be very useful as it is either a part of the grain that needs to be excluded to avoid lipid oxidation and rancidity, or a nutritionally interesting byproduct that could be followed during fractionation processes in order to get germ-rich fractions. The aim of this study was rst to improve the accuracy of the biochemical markers method by introducing new markers of testa and germ and by developing a better quantication of aleurone cell walls and cell contents, to evaluate the sensitivity of the method, and then to validate its use as a tool to assess the grain tissue proportions in technological fractions exhibiting very distinct compositions, obtained by different grain or bran fractionation processes. 2. Materials and methods 2.1. Preparation of grain tissues and fractions 2.1.1. Wheat samples Two common wheat (Triticum aestivum L.) cultivars differing in kernel hardness (hard wheat: cv. Tiger, and soft wheat: cv. Crousty), harvested in 2005 in Germany (Tiger) and France (Crousty) were used in this study. 2.1.2. Isolation of wheat grain tissues for biochemical analyses The structure and composition of the different wheat grain tissues have already been described (Barron et al., 2007; Hemery et al., 2007; Surget and Barron, 2005). To obtain isolated grain tissues, the grains ends (germ and brush) were removed with a razor blade and the remaining parts were immersed in distilled water for 1216 h at 20 C. A crease incision was made and the endosperm was removed using a scalpel. The crease area was removed and three different strips comprising different tissues were separated from these whole outer layers using a scalpel. Antoine et al. (2003) showed that the outer strip corresponds to the outer pericarp (epidermis and hypodermis), the inner strip corresponds to the aleurone layer, and the intermediate layer is a composite of several tissues (inner pericarp, testa and nucellar tissue). Scutellum and embryonic axis were also dissected. Nonadhering outer tissues surrounding the embryonic axis were removed and then the embryonic axis was separated from dry grains with a needle. The scutellum was removed with a scalpel after soaking the grain in distilled water for 1216 h at 20 C. These hand-isolated tissues were then used for the biochemical analyses. 2.1.3. Determination of the relative amount of tissues within the grain Similar hand dissections of 10 grains were performed in triplicate for each cultivar, to quantify the relative amounts of the different tissues, as described by Barron et al. (2007). All recovered tissues were weighed after drying at 25 C over phosphorus pentaoxide (P2O5). Outer layer tissues proportions were deduced from the combined weights of the aleurone, intermediate layer and outer pericarp, dissected from whole outer layers without the crease.
57
2.1.4. Production of the conventional milling fractions White our (0.55% ash content), whole meal our, ne bran and coarse bran fractions were produced using a conventional milling process, carried out on a Test-mill in the Department of Safety & Quality of Cereals, Federal Research Centre for Nutrition and Food (BFEL), Germany. 2.1.5. Production of bran and our fractions by peeling, pearling and milling Eight different bran and our fractions were provided by Buhler A.G. Uzwil, Switzerland. These fractions were obtained by two different debranning processes before milling (Fig. 1), by friction (peeling) and by abrasion (pearling), according to the published Buhler A.G. patent applications (Eugster and Gerschwiler, 2006). Peeling fractions were produced using a Buhler DC Peeler taking off 33.5% of the wheat kernel. Peeled wheat kernels were used as raw material for the pearling processing, using a Buhler stone polishing machine taking off approximately 3% more of the grain, resulting in pearled kernels and the pearling fractions. Peeled and pearled wheat kernels were tempered for 1215 h to a moisture content of 15.5%, and then milled on a Buhler laboratory mill to 76% extraction rate our (based on whole kernels before peeling and pearling) to give the 76% ours after peeling and after pearling, and the bran fractions after peeling and after pearling. For the production of reconstituted whole-grain our (100% extraction of the remaining kernel) from peeled and pearled kernels, coarse bran as well as bottom our were reduced to particle sizes <350 mm and were added and mixed to the our in the corresponding ratios. 2.1.6. Production of aleurone-rich fractions from bran Different aleurone-rich fractions were produced from Crousty and Tiger coarse brans by Buhler A.G. using physical dry processing. Details about the preparation of these aleurone fractions can be found in the Buhler A.G. patent applications (Bohm et al., 2003; Bohm and Kratzer, 2005). A rst bran fraction enriched in aleurone cells (Aleurone 1) was produced using grinding and air-classication, and sieved to give fractions above and below 180 mm. These Aleurone 1 >180 mm and Aleurone 1 <180 mm fractions were then further puried to give high purity fractions (Aleurone 2 > 180 mm and Aleurone 2 < 180 mm) using electrostatic separation. The Byproduct 1 and By-product 2 fractions correspond to the by-products of the Aleurone 1 and Aleurone 2 fractions, respectively.
2.2. Biochemical analyses Dissected tissues were dried at 25 C over P2O5, and then ground under cryogenic conditions for 4 min (with a Spex CertiPrep 6750 lab impact grinder), and dried again before chemical analyses. Milling fractions (except ours) were ground in a ball mill (Dangoumeau, France) for 4 min. 2.2.1. Isolation of aleurone cell wall material The amount of cell walls in Tiger and Crousty hand-isolated aleurone layers was assessed by gravimetric determination of insoluble cell wall material after proteolysis, using an adaptation of the method described by Brillouet et al. (1988). The dried and ground sample was suspended in hexane and centrifuged, the supernatant was discarded, these steps were repeated once, and the nal pellet was air dried. The dried pellet was suspended in water with 1.5% SDS and 5 mM Na-bisulte, a pronase solution was added (0.5 ml, 1 mg pronase/ml water), the mixture was agitated for 1 h at room temperature and then centrifuged. The pellet was washed extensively with water with intermittent centrifugation, and then successively suspended in 80% ethanol, absolute ethanol, acetone and ether, with centrifugation between two successive suspensions. The nal pellet corresponding to the aleurone cell wall material was dried at 25 C over P2O5 for 72 h, and weighed with 0.1 mg accuracy. This protocol was only used to evaluate the cell wall content in aleurone layers, and not for biochemical analyses. 2.2.2. Starch Total starch content of fractions was measured using Megazyme kits (Megazyme International Ireland Ltd., Ireland) according to approved AACC method 76-13 (AACC, 2000). Samples were analysed in duplicate with c.v. <10% (mean c.v. 4%). 2.2.3. Phytic acid The phytic acid content of ground tissues and milling fractions was measured at 500 nm from acidic extracts according to the colorimetric method of Latta and Eskin (1980) as modied by Vaintraub and Lapteva (1988). A standard curve was obtained with corn phytate (P-8810, Sigma) solutions of known concentrations. Samples were analysed in duplicate with coefcients of variation (c.v.) <10% (mean c.v. 6%).
Wheat grains
Peeling (friction)
Peeling fraction
Grains after peeling
Pearling (abrasion)
Milling
Milling
Pearling fraction
Grains after pearling
100% flour after peeling
76% flour after peeling
Milling
Milling
Bran after peeling
100% flour after pearling
76% flour after pearling
Bran after pearling

Fig. 1. Schematic representation of the whole-grain fractionation process, including peeling, pearling, and milling steps.
58
2.2.4. Phenolic acids Ester-linked phenolic acids were saponied under Argon (oxygen-free) at 35 C in 2 N sodium hydroxide. An internal standard (2,3,5-trimethoxy-(E)-cinnamic acid (TMCA), T-4002, Sigma Chemical Co., St Louis, USA) was added before adjusting pH to 2. Phenolic acids were then extracted with diethylether and quantied by RP-HPLC as described by Antoine et al. (2003). The response factors of the para-coumaric acid (p-CA) and the 4-O-80 , 50 -500 dehydrotriferulic acid (ferulic acid trimer, FAt) relative to the internal standard were determined at 320 nm with pure compounds. All analyses were performed at least in duplicate, with c.v. <10% (mean c.v. 3% for p-CA and 8% for FAt). 2.2.5. Alkylresorcinols Two different methods were used to quantify the total alkylresorcinol content of samples. For the calculation of grain tissues proportions, the quantication of each alkylresorcinol (C17:0 to C25:0) was carried out using gas chromatography, and the total alkylresorcinols in fractions from processed wheat grains and in hand-isolated tissues was determined as the sum of each identied molecule. Conditions of extraction, separation by gas chromatography and quantication of each alkylresorcinol (C17:0 to C25:0) were as described in Ross et al. (2003). Each determination was carried out in duplicate and all the c.v. were inferior to 10%. A colorimetric method was also used to measure the total amount of alkylresorcinols in some of the fractions. This colorimetric method was adapted from Tluscik et al. (1981) and used preferential coupling of alkylresorcinol to Fast Blue B reagent which is measured by changes in absorbance at 520 nm. Total amount of alkylresorcinol was estimated using a calibration curve (0100 mg/ml) prepared with olivetol (C5:0) as reference molecules. The most efcient solvent (n-propanol) and extraction conditions (two extractions at 20 C, 2 h each) were determined using preliminary studies. 2.2.6. Wheat germ agglutinin Wheat germ agglutinin (WGA) was determined in our and bran products by enzyme-linked immunosorbent assay (ELISA) as previously proposed by Vincenzi et al. (2002) with modications. WGA was extracted from samples with dilute hydrochloric acid, then extracts were serially diluted with phosphate buffer saline (PBS) containing bovine serum albumin (BSA). A commercial WGA was used to prepared calibration standards ranging from 0.2 to 125 ng/mL. Flat-bottomed microtiter plates were coated overnight with ovalbumin in PBS. After coating, plates were washed three times with PBS and saturated with 0.5% (w/v) BSA in PBS. Plates were washed again and incubated with calibration standards and serial dilutions of sample extracts. After a new washing step, plates were incubated with anti-WGA rabbit antibodies. Plates were washed 3 times and incubated with alkaline phosphatase conjugated puried anti-Rabbit immunoglobulins. After a last step of washing, the plates were incubated with enzyme substrate. Enzymatic reaction was stopped by addition of sodium hydroxide to the wells. Absorbances were read at 450 nm. A calibration curve was assessed with the standards, and WGA concentration in the extracts was computed by inversing the response function. 3. Results and discussion 3.1. Development of the method and evaluation of its sensitivity 3.1.1. Selection of the biochemical markers of wheat grain tissues Antoine et al. (2004) reported the use of starch, phytic acid, para-coumaric acid (p-CA) and a ferulic acid dehydrotrimer (FAt) as markers of endosperm, aleurone cell contents, aleurone cell walls, and outer pericarp, respectively. Starch is known to be
a characteristic of the starchy endosperm, and is thus a very good marker for this grain tissue. In wheat grain, phytic acid is known to be located in the aleurone cell contents and in the germ. Other compounds, like some B-vitamins, could also be chosen as markers of aleurone cell contents, but phytic acid presents the advantages of being very stable during dry-processes (unlike some heat-sensitive and storage-sensitive B-vitamins), and being easily quantied using simple methods. Antoine et al. (2004) used phytic acid as a marker of aleurone cell contents in bran fractions, as they assumed that coarse bran was free of germ. ODell et al. (1972) found that almost 90% of total phytic acid is located in the aleurone, the rest being in the germ. This repartition of phytic acid within the grain has been conrmed in the present study. Indeed, for Crousty cultivar, the embryonic axis, scutellum and aleurone layer contained 32.0, 67.2 and 156.3 mg/g phytic acid (on a dry matter basis) respectively, corresponding to 3%, 8% and 88% of the total phytic acid content of the grain. Phytic acid was thus chosen as marker of the aleurone cell contents. In the present study, ferulic acid trimer was found to be 11.5 and 22 times more concentrated in the outer pericarp than in the aleurone and in the intermediate layers respectively, and was therefore regarded as an appropriate marker of the outer pericarp (Table 1). The choice of para-coumaric acid as a marker was considered, according to the amounts quantied in the different grain tissues and in the whole outer layers. Antoine et al. (2004) and Barron et al. (2007) showed that p-CA is 25 times, 3 times, and >12 times more concentrated in the aleurone layer than in the outer pericarp, intermediate layer and germ (including scutellum). Thus, as p-CA is known to be bound to cell wall polysaccharides (Fincher and Stone, 1986; Grabber et al., 2004; Rhodes et al., 2002), this compound can be used as a marker of aleurone cell walls. However, as the amounts of p-CA found in the outer pericarp and the intermediate layer are not negligible, these amounts have to be taken into account and subtracted from the total p-CA content of a fraction in order not to bias the calculation of the proportion of aleurone cell walls in this fraction. Antoine et al. (2004) assumed that the sum of the outer pericarp, intermediate layer, aleurone and endosperm proportions in the studied fractions was equal to 100% (i.e. they supposed that the bran fractions did not contain germ), and then deduced the intermediate layer proportion by subtraction. If fractions containing germ were analysed using this method, the results may be biased and the intermediate layer proportion would be overestimated. A specic marker is thus needed to allow an accurate quantication of the intermediate layer proportion in fractions. Landberg et al. (2008)
Table 1 Composition of hand-isolated outer layers and peripheral tissues (reference values) Tissues Maximum c.v. Outer pericarp Intermediate layera Aleurone layer Endosperm Total outer layers Whole grains Cultivar FAt (mg/g) 10% 1.35 1.15 0.06 0.04 0.12 0.10 nd nd 0.34 0.28 0.04 0.03 Total ARs (mg/g) 10% 0.08 0.10 16.4 16.2 0.03 nd 0.02 0.01 4.07 3.93 0.47 0.42 p-CA (mg/g) 6% 0.01 0.02 0.11 0.09 0.29 0.22 nd nd 0.16 0.13 0.03 0.02 Phytic acid (mg/g) 9% 152.3 156.3 70.7 78.4 13.7 11.8 644.4 656.5 Starch (% w/w) 6% 78.3 78.1
Tiger Crousty Tiger Crousty Tiger Crousty Tiger Crousty Tiger Crousty Tiger Crousty
nd: non-detected; FAt: ferulic acid dehydrotrimer; ARs: alkylresorcinols; p-CA: paracoumaric acid. a Composed of the hyaline layer testa inner pericarp.
59
studied the location of alkylresorcinols in cereal grains and found that more than 99% of wheat grain total ARs are concentrated in the hand isolated so-called intermediate layer, and more precisely in the cuticle of the testa, showing that these phenolic lipids could be used as specic biochemical markers. The total amount of ARs was therefore chosen as a new marker of intermediate layer. Wheat germ agglutinin is a lectin found exclusively in wheat germ (Miller and Bowles, 1982). It was therefore selected as a potential marker of wheat germ in milling fractions and was used as a qualitative indicator to classify milling fractions according to their germ content. 3.1.2. Quantication of tissue proportions and validation by comparison with hand-dissection After selection of the biochemical markers and assessment of the biochemical composition of isolated tissues and fractions, the grain tissue proportions can be estimated. By combining the biochemical marker contents of hand-isolated grain tissues and the marker contents of the different technological fractions, the following relations allow calculation of the grain tissue proportions within fractions.
. %aw p-CAF %pp-CAp %ip-CAi p-CAaw 100 %ac Phytic acidF =Phytic acidac 100 %e StarchF =Starche 100
with:
%p FAtF =FAtp 100 %i ARsF =ARsi 100
p-CAaw p-CAa =awa Phytic acidac Phytic acida =aca

where: (%p), (%i), (%ac), (%aw), and (%e): proportion of outer pericarp, intermediate layer, aleurone cell content, aleurone cell walls, and starchy endosperm in studied fractions. [FAt]: ferulic acid dehydrotrimer content; [ARs]: total alkylresorcinols content; [p-CA]: para-coumaric acid content; [phytic acid]: phytic acid content; [starch]: starch content, in studied fractions (F), outer pericarp (p), intermediate layer (i), total aleurone (a), aleurone cell walls (aw), aleurone cell content (ac), and starchy endosperm (e). [aw]a and [ac]a: concentration of cell walls and cell contents in the aleurone layer. The approximate proportions of germ in fractions (%g) was calculated as follows:
microscopy observations carried out in previous studies (Antoine et al., 2003; Bacic and Stone, 1981; Evers and Reed, 1988; Rhodes et al., 2002), that showed thick aleurone cell walls accounting for an important proportion of aleurone cell total volume. To test the efciency of the described tissue proportion calculation method, the composition of whole grains and whole outer layers was calculated and compared to the composition assessed by hand dissection and weighing. The comparison of the results obtained for Crousty and Tiger grains by both methods (markers and dissection) is shown in Table 2. Almost no difference was found between the endosperm amounts, and the total outer layer proportions obtained by both methods were very close (0.6% and 0.1% absolute error for Crousty and Tiger grains respectively). The proportions of peripheral tissues were found to be a little more variable (from 0.4% to 1.4% absolute error) depending on the method used, but these variations in peripheral tissue proportions are due to the fact that these tissues are present in small amounts in the ground grains analysed, and that biochemical analyses are less precise on small amounts. These results show that the biochemical markers method allows the distinct quantication of tissues proportions in wheat fractions that contain very few peripheral layers, with an absolute error of 1.5%. About 2.5% of the wheat grain was not quantied with this method. This proportion is comparable to the proportion of germ found in wheat grain by hand-dissection. Fig. 2 presents the tissue proportions obtained for Crousty and Tiger whole outer layers. The tissues proportions calculated using the biochemical markers are very close to those assessed by hand dissection. The mean relative errors are 4% and 8% for Crousty and Tiger respectively, and the total of the tissue proportions are 98% and 96%. These results show that this method allows effective quantication, and suggest that it may be very well adapted for the assessment of grain tissue proportions in fractions rich in bran tissues. 3.1.3. Impact of analytical variability and possible simplication of the method This method is based on the measurement of 5 different compounds (FAt, ARs, p-CA, phytic acid and starch) on two sample sets (i.e. the hand-isolated reference tissues and the technological fractions), it implies therefore 10 quantitative determinations. The random analytical error on the tissue proportions was calculated for Tiger and Crousty coarse bran, by taking into account the reproducibility (i.e. the mean c.v. of the dosage) of each measurement. Coarse bran was chosen for the calculation of variability because this fraction contains all of the grain tissues. If these calculations had been carried out for other samples, the results could have been different as the random analytical error depends on the composition of the studied fraction, due to the fact that the reproducibility varies from one dosage to another. The effects on
%g WGAF =WGAgrain germgrain

where [WGA]F and [WGA]grain are the WGA content in fractions and ground whole grain respectively, and [germ]grain is the proportion of germ (scutellum plus embryonic axis) obtained by hand dissection of grains. In order to separately quantify the proportions of aleurone cell walls and cell contents in the different fractions, the results have to be calculated in relation to the amount of markers found in aleurone cell walls and aleurone cell contents (expressed in mgp-CA/ galeurone cell walls and in mgphytates/galeurone cell contents), and not in relation to the amount of markers found in the whole aleurone layer (expressed in mgmarker/galeurone), otherwise these proportions would be over-estimated. Thus, the amount of aleurone cell walls was assessed by gravimetric determination of insoluble cell wall material. Crousty and Tiger aleurone layers were found to contain 42.4% 2.5% and 48.0% 2.6% cell walls, respectively. This high proportion of cell walls in the aleurone layer corresponds to
Table 2 Comparison of Crousty and Tiger grains composition (%), using either biochemical markers or hand-dissection Grains tissues Crousty Markers Total outer layers Outer pericarp Intermediate layer Aleurone cell walls Aleurone cell contents Starchy endosperm Embryo scutellum Total of grain tissues Unknown fraction 13.7 2.7 2.6 4.1 4.4 84.1 97.8 2.2 Dissection 13.1 3.5 3.2 2.7 3.7 84.1 2.8 100 Tiger Markers 14.7 3.1 2.9 4.1 4.7 82.4 97.1 2.9 Dissection 14.6 4.3 3.3 3.4 3.6 82.7 2.7 100
The unknown fractions is the difference between 100% and the sum of all the grain tissues proportions.
60
100%
98% 29% 21% 24%
100% 28% 21% 24%
96% 24% 22% 25%
100% 25% 23% 22%
immunological tests in order to further simplify the determination of the biochemical markers. 3.2. Biochemical markers as a tool to monitor fractionation processes The method was then used to assess the histological proportion of fractions exhibiting very different compositions, obtained by three different grain or bran dry fractionation processes, in order to verify if this method effectively allows study of the distribution of the different tissues among fractions, and to evaluate if this method can be a useful tool. 3.2.1. Conventional milling process: whole-grain fractionation The calculated grain tissue composition of whole grains, whole meal ours, white ours, ne brans and coarse brans is shown in Table 3. The composition of the fractions made of both cultivars was quite similar, except for ne brans. Indeed, Tiger ne bran contains twice as much endosperm as Crousty ne bran, may be due to differences in bran brushing steps. Whole grain and whole meal our exhibit very similar composition, and the white our (0.55% ash content) contains only 12% of non-endosperm tissues. For most of the fractions, the sum of the tissue proportion is superior to 85%, but for ne brans, totals of 73% and 67% are observed for Tiger and Crousty respectively, suggesting that a non-negligible part of these fractions was not quantied. In conventional milling processes, wheat germs are often found in high amounts in the ne bran fractions (Posner and Li, 1991). Fig. 3 shows that there is a relation between the approximate germ proportion and the nonquantied proportion in fractions, suggesting that the non-quantied 30% in ne brans may effectively be composed of germ. Fig. 4 presents the proportions of each of the peripheral layers relative to the total amount of outer layers in bran fractions. Crousty and Tiger coarse brans exhibit tissue proportions similar to those found for whole hand-isolated outer layers, except for the outer pericarp proportion (about 19% outer pericarp in coarse bran compared to 25% in hand-isolated outer layers). This difference may be due to the high friability of outer pericarp, which is a brittle material that can be easily removed from the other layers during milling and fragmented into ner particles not recovered in coarse brans (Antoine et al., 2003, 2004). Fig. 4 also shows that the proportions of cell walls in the total aleurone layer were almost the same for coarse and ne bran as for whole hand-isolated outer layers, showing that the aleurone cell walls and cell contents were not dissociated in these fractions (i.e. the aleurone cells were not damaged).
24% 0% Markers
27%
25%
30%
Dissection
Markers
Dissection
CROUSTY
outer pericarp aleurone cell walls
TIGER
intermediate layer aleurone cell contents
Fig. 2. Comparison of Crousty and Tiger outer layers composition (%), using either biochemical markers or hand-dissection.
under-estimates or over-estimates of the biochemical markers content in reference hand-isolated tissues and/or in coarse bran were calculated. The mean, low (mean value 1 standard deviation) and high (mean value 1 standard deviation) values of each biochemical marker were calculated for each grain tissue and for coarse bran. Then, the grain tissue proportions resulting from the combinations of all these values were calculated. The lowest variability was observed for the aleurone cell walls (4%) and the highest for the intermediate layer (22%). The maximum relative error (i.e. random analytical error) observed for coarse bran was 13% (similar results were obtained for both cultivars). The impact of analytical variability on tissue proportion calculation can thus be considered as acceptable. To determine the tissue distribution in fractions obtained from wheat grain processing, four different biochemical assays need to be carried out, i.e. a rapid and simple colorimetric method to estimate the phytic acid content, an enzymatic assay to measure starch content which is standardized as a commercial kit, and two more complex and time consuming assays to analyse either the phenolic acids or the alkylresorcinols composition and content. However, if analysis of the composition in different phenolic acids appeared necessary in order to distinguish between distinct tissues as the pericarp and the aleurone layer, only the total content of alkylresorcinols is required as all of these compounds were found to be concentrated in the intermediate layer of wheat grains (Landberg et al., 2008). Thus, in order to simplify and reduce the time for biochemical analyses, we tested the potential of a colorimetric method for the determination of alkylresorcinols using Fast Blue B reagent as described by Tluscik et al. (1981). Choice of the most effective solvent and of the extraction conditions were adapted to fractionation samples after preliminary studies (cf. Section 2). A very good correlation was observed between determination of total AR content using either the colorimetric method or the sum of each of the distinct AR molecules separated by gas chromatography (colorimetric method 0.91 GC method, R2 0.98). This result agrees with previous studies (Andersson et al., in press) and shows that it is possible to develop a simpler and quicker assay for determination of wheat grain tissue proportions in fractions obtained from processing. Therefore, only the analysis of phenolic acids appears as the limiting factor when trying to simplify the assays. However it seems possible, in future, to develop specic
Table 3 Grain tissues proportions (%) in fractions produced by a conventional milling process (Tiger and Crousty cv.) Outer Intermediate Aleurone layer Starchy Total pericarp layer endosperm Total Cell Cell walls content Whole grain Whole meal our White our Fine bran Coarse bran Tiger 3.1 Crousty 2.7 Tiger 3.0 Crousty 2.4 Tiger 0.1 Crousty 0.2 Tiger 6.1 Crousty 7.1 Tiger 15.9 Crousty 15.1 2.9 2.6 3.2 2.7 0.1 0.1 6.6 8.0 23.0 21.9 8.8 8.5 8.1 8.5 1.2 1.5 4.1 4.1 3.9 3.7 0.5 0.5 4.7 4.4 4.2 4.8 0.7 1.0 9.5 14.0 21.6 24.9 82.4 84.1 69.9 75.0 99.9 97.0 42.9 25.3 8.0 8.0 97 98 84 89 101 99 73 67 89 91
17.4 7.9 26.6 12.6 42.5 20.9 46.3 21.3
61
45 y = 0,81x 40 35 R2 = 0,74
30 25 20 15 10 5 0 0 5 10 15 20 25 30 35 40 45
Unknown proportion (%)

Fig. 3. Comparison of the approximate germ proportions with the unknown proportions in milling fractions. The unknown proportion in a fraction is the difference between 100% and the sum of all the calculated grain tissues proportions.
3.2.2. Whole-grain fractionation by peeling, pearling, and milling This process, by sequential removal of grain outer layers by debranning prior to milling, aims at producing less contaminated wheat grains and whole-grain ours, by removing the outermost layers of the grain. The rst debranning step carried out by friction removes 34% of the grain by weight. As shown in Table 4, this removed fraction (i.e. the peeling fraction) contains 53% and 60% of outer pericarp for Tiger and Crousty respectively. It also contains various proportions of all the other grain tissues, including endosperm, probably due to the presence of broken grains in the peeling fractions and to the fact that the grain ends are more worn away during debranning. This rst debranning step by peeling removes the major part of the outer pericarp of the grain (up to 60%). Due to the shape of wheat grains, removing all the outer pericarp by peeling does not seem possible as about 25% of this tissue is located in the crease (Evers and Millar, 2002). The pearling fractions, that correspond to the fraction removed off the grain during the second debranning step (about 3% additional debranning), contains
3035% aleurone material and 2530% starchy endosperm, showing that all the peripheral layers have been removed on some parts of the grains: the grain ends are more eroded and the grains are rounded by pearling. These pearling fractions also contain 713% of germ. As a large part of the outer pericarp was removed during the friction step, the other bran fractions and the our fractions obtained afterwards do not contain much of this tissue (Table 4). As a consequence, the 100% ours from peeling contain less outer layer particles than the whole meal ours, and the 100% ours from pearling, which are made from grains that have undergone an additional debranning step (pearling), contain even less outer layers. All these 100% ours contain 34% germ, that is close to the amount of germ quantied in whole grains (Table 2). The 76% ours from peeling and pearling are both almost only composed of starchy endosperm (only 2% of non-endosperm particles). Fig. 4 presents the proportions of each of the outer layers relative to the total amount of peripheral layers in bran fractions. The pearling fractions contain the most important proportion of intermediate layer (810% more than in hand-isolated whole outer layers), due to the removal of outer pericarp by peeling and to the fact that a part of the aleurone layer probably remains attached to the grains and is found in the bran fraction after pearling. Surprisingly, bran fractions after peeling and after pearling exhibit quite similar tissues proportions. The differences between these two fractions are probably reduced because they both contain the bran present in the crease area. This crease bran represents a nonnegligible proportion of the grain total outer layers and is not easily removed neither by peeling nor by pearling (Dexter and Wood, 1996). Fig. 4. shows that for almost all the bran fractions, the ratio aleurone cell walls/aleurone cell contents is very close to that observed for coarse bran and hand-isolated whole outer layers, suggesting that aleurone material is not dissociated and that the aleurone cells are not broken. Only the peeling fractions exhibit a surprisingly high proportion of aleurone cell walls compared to their proportion of aleurone cell contents. This may be explained by the fact that the cell walls, being linked to the other outer layers, are torn off the grain by peeling, leaving the cell contents on the grain. But this high proportion of aleurone cell walls in peeling fractions seems quite excessive when compared to the proportions of intermediate layers in these fractions, suggesting that it may have been over-estimated due to the high content of para-coumaric acid in these fractions. If the proportions of total aleurone layer is calculated by taking into account only the phytic acid content and not the p-CA content of the peeling fractions, the total aleurone content in these fractions is found to be 34%. This could be more
Germ proportion (%)
25 23 26 26 Tg
29 22 25 24 Cr
31 26 22 20 Tg
34 30 19 17 Cr
27 26 28 20 Tg
30 26 26 18 Cr
26 8
27 4
32 26
38 26 30 6 Cr
29 24 35 11 Tg
27 27 34 11 Cr
33 27 30 10 Tg
35 29 28 8 Cr
63
68
32 9
Tg
Cr
Tg
Hand-isolated whole bran
Fine bran
Coarse bran
peeling fraction
bran fraction after peeling
pearling fraction aleurone cell contents
bran fraction after pearling
outer pericarp
intermediate layer
aleurone cell walls
Fig. 4. Study of the dissociation of the outer layers in the different Tiger and Crousty bran fractions: histological composition of the particles of outer tissues present in these fractions (i.e. endosperm excluded). The proportions of outer pericarp, intermediate layer, and aleurone layer are re-calculated relatively to the total amount of these tissues in bran fractions.
62
Table 4 Grain tissues proportions (%) in fractions produced by a whole-grain fractionation process including peeling, pearling, and milling steps (Tiger and Crousty cv.) Outer pericarp Intermediate layer Aleurone layer Total Peeling fraction 100% our from peeling 76% our from peeling Bran fraction after peeling Pearling fraction 100% our from pearling 76% our from pearling Bran fraction after pearling Tiger Crousty Tiger Crousty Tiger Crousty Tiger Crousty Tiger Crousty Tiger Crousty Tiger Crousty Tiger Crousty 53.2 59.5 1.4 1.0 0.1 0.2 5.9 4.9 6.9 7.0 1.1 0.8 0.2 0.1 6.3 5.0 7.2 3.4 3.3 3.0 0.3 0.2 20.9 23.2 21.4 21.5 2.5 2.0 0.5 0.5 19.4 17.2 24.4 24.9 7.2 7.9 1.1 1.4 38.0 48.9 32.1 33.8 5.6 5.6 1.5 1.6 38.5 39.9 Cell walls 22.2 23.4 2.8 2.9 0.2 0.3 16.9 20.0 14.8 16.8 2.4 2.4 0.3 0.2 17.2 17.9 Cell content 2.2 1.5 4.4 4.9 0.8 1.1 21.0 28.9 17.4 16.9 3.2 3.2 1.2 1.3 21.3 21.9 7.5 9.5 83.8 86.2 97.5 99.9 22.4 16.1 26.2 29.5 85.1 89.8 102.5 100.4 22.7 21.9 92 97 96 98 99 102 87 93 87 92 94 98 105 103 87 84 Starchy endosperm Total
plausible, as these fractions are also supposed to contain 917% of germ. The high amount of p-CA quantied in the peeling fractions might be due to an external contamination. Indeed, contaminating materials such as straw and chaff can be found in the debranning fraction, and several authors showed that wheat straw is very rich in para-coumaric acid (Benoit et al., 2006; Sun et al., 2001; Tapin et al., 2006). 3.2.3. Bran fractionation process: production of aleurone-rich fractions The wheat aleurone layer has been shown to have great nutritional interest, and to concentrate most of the micronutrients (minerals and vitamins), antioxidants (mainly phenolic acids) and other phytochemicals of the wheat grain (Hemery et al., 2007). Thus, it can be interesting to produce bran fractions highly concentrated in aleurone material, to transform a by-product of our into a high nutritional value food ingredient. In the present study, various aleurone-rich bran fractions were studied (Table 5). The Aleurone 2 <180 mm, which is the most concentrated fraction, contains 79% and 89% aleurone, while coarse bran contains only 43% and 46% aleurone for Tiger and Crousty respectively. In the aleurone-rich fractions, the proportion of aleurone material increases from 55% to 79% and from 65% to 89% for Tiger and
Crousty respectively. This difference between Tiger and Crousty cultivars may be explained by the fact that Crousty peripheral layers contain 4% more aleurone material than Tiger ones (Fig. 2). Thus, Crousty bran fractions are naturally richer in aleurone material. This higher proportion of aleurone material in the fractions of Crousty (soft wheat) than in the fractions of Tiger (hard wheat) can also be explained by the distinct behaviors of soft and hard common wheats at break milling stages, which result in different repartitions of the aleurone cell contents between starchy endosperm and bran, as already pointed out by Greffeuille et al. (2005). Table 5 also shows that ner particles are richer in aleurone materials. The biggest particles which are less dissociated are more likely to be composite particles made of different tissues. The second step of the process (giving Aleurone 2) increases the aleurone content of the products. In parallel to the increase in aleurone proportion in fractions, the outer pericarp proportion decreases from 23% to 8% and from 19% to 9% for Tiger and Crousty respectively, this outer pericarp being discarded in the by-products. The by-products 1 are richer in intermediate layer than the by-products 2, but in the case of Crousty cultivar, these two fractions are still very similar. For Tiger cultivar, an increase in outer pericarp proportion is observed in by-product 2 (compared to by-product 1), as a result of the decrease of outer pericarp content in Aleurone 2
Table 5 Grain tissues proportions (%) in fractions produced by the Buhler A.G. LEURON bran fractionation process (Tiger and Crousty cv.) Outer pericarp Intermediate layer Aleurone layer Total Aleurone 1 >180 mm Aleurone 1 <180 mm Aleurone 2 > 180 mm Aleurone 2 < 180 mm By-product 1 By-product 2 Tiger Crousty Tiger Crousty Tiger Crousty Tiger Crousty Tiger Crousty Tiger Crousty 22.9 19.1 17.8 16.9 9.1 11.2 7.8 8.5 30.9 36.6 45.6 33.2 13.6 13.1 11.8 12.6 12.9 12.7 11.1 12.0 18.9 18.6 11.0 12.7 55.3 64.7 64.3 70.9 74.2 83.2 79.3 89.1 29.7 29.4 27.6 32.9 Cell walls 28.3 31.0 32.1 34.7 32.7 38.9 37.3 42.4 20.5 19.9 15.9 20.2 Cell content 27.0 33.7 32.2 36.2 41.5 44.3 42.0 46.7 9.2 9.6 11.7 12.7 2.0 2.4 2.3 2.5 2.6 2.7 2.9 3.0 3.6 4.4 1.6 2.0 94 99 96 103 99 110 101 113 83 89 86 81 Starchy endosperm Total
63
fractions (compared to Aleurone 1 fractions) after the second fractionation step. For almost all the aleurone fractions, the ratio aleurone cell walls/aleurone cell contents is really close to the ratio observed for hand-isolated aleurone layers, suggesting that aleurone material is not dissociated (i.e. that the aleurone cells are not broken). This observation are in agreement with microscopy observations (Amrein et al., 2003; Buri et al., 2004) showing that these aleuronerich fractions are composed of clusters of intact cells. All these observations are in agreement with the observations made by several authors. Indeed, the higher protein and beta-glucan contents observed by Amrein et al. (2003), the higher phenolic acids content observed by Zhou et al. (2004), and the higher minerals and phytic acid content observed by Buri et al. (2004) for Aleurone 2 than for Aleurone 1, can be explained by the enrichment in aleurone material of Aleurone 2 compared to Aleurone 1.
dissection, and to apply this quantication method to batches of grains of various origins. Several biochemical analyses have to be carried out to assess the composition of fractions using the described method. Therefore, it would be difcult to use it for controls on the production line, but this method can be useful in quality control laboratories of rms, and can also be used as a reference for the calibration of equipment based on physical properties (i.e. uorescence). This biochemical marker method has been proved to be an efcient tool for assessment of grain tissue proportions in various milling fractions exhibiting contrasting composition, such as ours, brans, or aleurone-rich fractions. The present study showed that this quantitative method allows, by calculation of the composition of the different products generated by grain or bran fractionation processes, an understanding of the effects of these processes and monitoring and adapting of them to reach the objective (i.e. efcient debranning, production of specic fractions). Acknowledgements The authors thank the Federal Research Centre for Nutrition and Food (BFEL), Germany, for the production of the conventional milling fractions used in this study. We also gratefully thank T.-M. Lasserre for her work on the biochemical analyses, and S. Chay and A. Putois for the manual dissection of grain tissues. This publication is nancially supported by the European Commission in the Communities 6th Framework Program, Project HEALTHGRAIN (FOOD-CT-2005-514008). It reects the authors views and the Community is not liable for any use that may be made of the information contained in this publication. References
AACC, 2000. American Association of Cereal Chemists Approved Methods, 10th ed. The Association, St Paul, Minnesota. ` Amrein, T.M., Granicher, P., Arrigon, E., Amado, R., 2003. In vitro digestibility and colonic fermentability of aleurone isolated from wheat bran. LebensmittelWissenschaft und Technology 36, 451460. Andersson, A.M.A., Kamal-Eldin, A., Fras, A., Boros, D., man, P. Alkylresocinols in wheat varieties in the HEALTHGRAIN diversity screen. Journal of Agricultural and Food Chemistry, in press. Antoine, C., Peyron, S., Mabille, F., Lapierre, C., Bouchet, B., Abecassis, J., Rouau, X., 2003. Individual contribution of grain outer layers and their cell wall structure to the mechanical properties of wheat bran. Journal of Agricultural and Food Chemistry 51, 20262033. Antoine, C., Peyron, S., Lullien-Pellerin, V., Abecassis, J., Rouau, X., 2004. Wheat bran tissue fractionation using biochemical markers. Journal of Cereal Science 39, 387393. Aureli, G., DEgidio, M.G., 2007. Efcacy of debranning on lowering of deoxynivalenol (DON) level in manufacturing processes of durum wheat. Tecnica Molitoria 58, 729733. Bach Knudsen, K.E., Steenfeldt, S., Borsting, C.F., Eggum, B.O., 1995. The nutritive value of decorticated mill fractions of wheat. 1. Chemical composition of raw and enzyme treated fractions and balance experiments with rats. Animal Feed Science and Technology 52, 205225. Bacic, A., Stone, B.A., 1981. Isolation and ultrastructure of aleurone cell walls from wheat and barley. Australian Journal of Plant Physiology 8, 453474. Baldwin, P.M., Bertrand, D., Novales, B., Bouchet, B., Collobert, G., Gallant, D.J., 1997. Chemometric labeling of cereal tissues in multichannel uorescence microscopy images using discriminant analysis. Analytical Chemistry 69, 43394348. Barrier-Guillot, B., Casado, P., Maupetit, P., Jondreville, C., Gatel, F., Larbier, M., 1996. Wheat phosphorus availability. 1. In vitro study; factors affecting endogenous phytasic activity and phytic phosphorus content. Journal of the Science of Food and Agriculture 70, 6268. Barron, C., Surget, A., Rouau, X., 2007. Relative amounts of tissues in mature wheat (Triticum aestivum L.) grain and their carbohydrate and phenolic acid composition. Journal of Cereal Science 45, 8896. Benoit, I., Navarro, D., Marnet, N., Rakotomanomana, N., Lesage-Meessen, L., Sigoillot, J.C., Asther, M., Asther, M., 2006. Feruloyl esterases as a tool for the release of phenolic compounds from agro-industrial by-products. Carbohydrate Research 341, 18201827. Bohm, A., Bogoni, C., Behrens, R., Otto, T., 2003. Method for the extraction of aleurone from bran. US Patent Application Publication US 2003/0175384 A1. Applicant: Buhler AG. Bohm, A., Kratzer, A., 2005. Method for isolating aleurone particles. US Patent Application Publication US 2005/0103907 A1. Applicant: Buhler AG.
4. Conclusion The evaluation of the method described here, in comparison with tissue quantication by hand dissection, has shown its efcacy (mean relative errors of 4% and 8% for Crousty and Tiger outer layers respectively), and the impact of the analytical variability (maximum 13% relative error on coarse bran) was regarded as acceptable. Some improvements can be made to the current method. For example, a better quantication of the germ content is needed, as the germ proportion is often not negligible in bran factions and fractions coming from debranning processes. Wheat germ agglutinin seems to be a promising marker of germ. The quantication method does not currently enable the accurate quantication of the germ proportions in milling fractions, but it gives qualitative information and allows classication of these fractions according to their germ content (Fig. 3). This method should be improved to add WGA as another marker to complete the current method. In the milling industry, batches of different grain cultivars are frequently used as starting material for dry fractionation processes. In this study, the tissue proportions have been deduced from reference values assessed on exactly the same wheat cultivars from which the fractions were produced: the genetic and agronomic variability was thus not taken into account. Some of the constituents chosen as biochemical markers have specic functions in wheat grain. Therefore, the natural variability of compounds such as phytic acid, phenolic acids, and alkylresorcinols among a wheat population could be very high. For example, Barrier-Guillot et al. (1996) showed that the amount of Phytic P can vary from 0.92 to 2.8 mg/g dm for common wheat with various agronomic and genetic characteristics. Within the European HEALTHGRAIN project (Poutanen et al., 2008), other authors analysed the numerous wheat varieties in the diversity screen of the project and showed that the total phenolic acid content varies from 326 to 1171 mg/g dm in the 130 winter wheat lines and from 456 to 892 mg/g dm in the 20 spring wheat lines (Li et al., in press), while the total alkylresorcinol content varies from 220 to 652 mg/g dm in these winter wheat lines and from 254 to 537 mg/g dm in the spring wheat lines (Andersson et al., in press). To try to evaluate the impact of genetic variability on grain tissue proportions using biochemical markers, tissue proportions of Tiger coarse bran were calculated with Crousty reference values (and vice versa). The mean relative error observed was 10%, and this variability was found to be less than the maximum analytical variability. This study should be carried out on other wheat cultivars. If this observation were conrmed, it could then be possible to use standard reference values (i.e. the means of the values obtained for the grain tissues of different cultivars) to quantify tissues proportion in fractions, to limit grain tissues
64
Y. Hemery et al. / Journal of Cereal Science 49 (2009) 5564 Li, L., Shewry, P.R., Ward, J.L. Phenolic acids in wheat varieties in the HEALTHGRAIN diversity screen. Journal of Agricultural and Food Chemistry, in press. Liu, R.H., 2007. Whole grain phytochemicals and health. Journal of Cereal Science 46, 207219. Mateo Anson, N., Van den Berg, R., Havenaar, R., Bast, A., Haenen, G., 2008. Ferulic acid from aleurone determines the antioxidant potency of wheat grain (Triticum aestivum L.). Journal of Agricultural and Food Chemistry 56, 55895594. Miller, R.C., Bowles, D.J., 1982. A comparative-study of the localization of wheatgerm-agglutinin and its potential receptors in wheat grains. Biochemical Journal 206, 571576. ODell, B.L., De Boland, A.R., Koirtyohann, S.R., 1972. Distribution of phytate and nutritionally important elements among the morphological components of cereal grains. Journal of Agricultural and Food Chemistry 20, 718723. Peyron, S., Surget, A., Mabille, F., Autran, J.C., Rouau, X., Abecassis, J., 2002. Evaluation of tissue dissociation of durum wheat grain (Triticum durum desf.) generated by the milling process. Journal of Cereal Science 36, 199208. Pomeranz, Y., 1988. Chemical composition of kernel structures. In: Pomeranz, Y. (Ed.), Wheat: Chemistry and Technoloy, vol. 4. AACC, St Paul, pp. 97158. Posner, E.S., Li, Y.Z., 1991. A technique for separation of wheat germ by impacting and subsequent grinding. Journal of Cereal Science 13, 4970. Poutanen, K., Shepherd, R., Shewry, P., Delcour, J.A., Bjorck, I., van der Kamp, J.W., 2008. Beyond whole grain: the European HEALTHGRAIN project aims at healthier cereal foods. Cereal Foods World 53, 3235. Pussayanawin, V., Wetzel, D.L., Fulcher, R.G., 1988. Fluorescence detection and measurement of ferulic acid in wheat milling fractions by microscopy and HPLC. Journal of Agricultural and Food Chemistry 36, 515520. Rhodes, D.I., Sadek, M., Stone, B.A., 2002. Hydroxycinnamic acids in walls of wheat Aleurone cells. Journal of Cereal Science 36, 6781. Ross, A.B., Shepherd, M.J., Schupphaus, M., Sinclair, V., Alfaro, B., Kamal-Eldin, A., Aman, P., 2003. Alkylresorcinols in cereals and cereal products. Journal of Agricultural and Food Chemistry 51, 41114118. Sun, R.C., Sun, X.F., Zhang, S.H., 2001. Quantitative determination of hydroxycinnamic acids in wheat, rice, rye, and barley straws, maize stems, oil palm frond ber, and fast-growing poplar wood. Journal of Agricultural and Food Chemistry 49, 51225129. Surget, A., Barron, C., 2005. Histologie du grain de ble. Industrie des Cereales 145, 37. Symons, S.J., Dexter, J.E., 1996. Aleurone and pericarp uorescence as estimators of mill stream renement for various canadian wheat classes. Journal of Cereal Science 23, 7383. Tapin, S., Sigoillot, J.C., Asther, M., Petit-Conil, M., 2006. Feruloyl esterase utilization for simultaneous processing of nonwood plants into phenolic compounds and pulp bers. Journal of Agricultural and Food Chemistry 54, 36973703. Tluscik, F., Kozubek, A., Mejbaum-Katzenellenbogen, W., 1981. Alkylresorcinols in rye (Secale-Cereale L) grains. 6. Colorimetric micromethod for the determination of alkylresorcinols with the use of diazonium salt, fast blue-B. Acta Societatis Botanicorum Poloniae 50, 645651. Vaintraub, I.A., Lapteva, N.A., 1988. Colorimetric determination of phytate in unpuried extracts of seeds and the products of their processing. Analytical Biochemistry 175, 227230. Vincenzi, S., Zoccatelli, G., Perbellini, F., Rizzi, C., Chignola, R., Curioni, A., Peruffo, A.D.B., 2002. Quantitative determination of dietary lectin activities by enzyme-linked immunosorbent assay using specic glycoproteins immobilized on microtiter plates. Journal of Agricultural and Food Chemistry 50, 62666270. Zhou, K., Laux, J.J., Yu, L., 2004. Comparison of swiss red wheat grain and fractions for their antioxidant properties. Journal of Agricultural and Food Chemistry 52, 11181123.
Brillouet, J.M., Rouau, X., Hoebler, C., Barry, J.L., Carre, B., Lorta, E., 1988. A new method for determination of insoluble cell-walls and soluble nonstarchy polysaccharides from plant materials. Journal of Agricultural and Food Chemistry 36, 969979. Buri, R.C., von Reding, W., Gavin, M.H., 2004. Description and characterization of wheat aleurone. Cereal Foods World 49, 274282. Chen, Y., Ross, A.B., Aman, P., Kamal-Eldin, A., 2004. Alkylresorcinols as markers of whole grain wheat and rye in cereal products. Journal of Agricultural and Food Chemistry 52, 82428246. Courcoux, P., Devaux, M.F., Bouchet, B., 2002. Simultaneous decomposition of multivariate images using three-way data analysis. Application to the comparison of cereal grains by congocal laser scanning microscopy. Chemometrics and Intelligent Laboratory Systems 62, 103113. Dexter, J.E., Wood, P.J., 1996. Recent applications of debranning of wheat before milling. Trends in Food Science & Technology 7, 3541. Eugster, W., Gerschwiler, O., 2006. Method and installation for cleaning cereal. US Patent Application Publication US 2006/0147591 A1. Applicant: Buhler AG. Evers, A.D., Reed, M., 1988. Some novel observations by scanning electron microscopy on the seed coat and nucellus of the mature wheat grain. Cereal Chemistry 65, 8185. Evers, A.D., Millar, S., 2002. Cereal grain structure and development: some implications for quality. Journal of Cereal Science 36, 261284. Fincher, G.B., Stone, B.A., 1986. Cell walls and their components in cereal grain technology. In: Pomeranz, Y. (Ed.), Advances in Cereal Science and Technology, vol. 8. American Association of Cereal Chemists, St Paul, MN, pp. 207295. Fleurat-Lessard, F., Chaurand, M., Marchegay, G., Abecassis, J., 2007. Effects of processing on the distribution of pirimiphos-methyl residues in milling fractions of durum wheat. Journal of Stored Products Research 43, 384395. Fulcher, R.G., OBrien, T.P., Lee, J.W., 1972. Studies on the aleurone layer. I. Conventional and uorescence microscopy of the cell wall with emphasis on phenol-carbohydrate complexes in wheat. Australian Journal of Biological Science 25, 2334. Grabber, J.H., Ralph, J., Lapierre, C., Barriere, Y., 2004. Genetic and molecular basis of grass cell-wall degradability. I. Lignin-cell wall matrix interactions. Comptes Rendus Biologies 327, 455465. Greffeuille, V., Abecassis, J., Bar LHelgouach, C., Lullien-Pellerin, V., 2005. Differences in the aleurone layer fate between hard and soft common wheats at grain milling. Cereal Chemistry 82, 138143. Hemery, Y., Rouau, X., Lullien-Pellerin, V., Barron, C., Abecassis, J., 2007. Dry processes to develop wheat fractions and products with enhanced nutritional quality. Journal of Cereal Science 46, 327347. Jacobs, D.R., Steffen, L.M., 2003. Nutrients, foods, and dietary patterns as exposures in research: a framework for food synergy. American Journal of Clinical Nutrition 78, 508S513S. Jensen, S.A., Munck, L., Martens, H., 1982. The botanical constituents of wheat and wheat milling fractions. 1. Quantication by autouorescence. Cereal Chemistry 59, 477484. Jensen, S.A., Martens, H., 1983. The botanical constituents of wheat and wheat milling fractions. 2. Quantication by amino-acids. Cereal Chemistry 60, 172177. Laca, A., Mousia, Z., Diaz, M., Webb, C., Pandiella, S.S., 2006. Distribution of microbial contamination within cereal grains. Journal of Food Engineering 72, 332338. Landberg, R., Kamal-Eldin, A., Salmenkallio-Marttila, M., Rouau, X., Aman, P., 2008. Localization of alkylresorcinols in barley, wheat and rye kernels. Journal of Cereal Science 48, 401406. Latta, M., Eskin, M., 1980. A simple and rapid colorimetric method for phytate determination. Journal of Agricultural and Food Chemistry 28, 13131315. Lempereur, I., Surget, A., Rouau, X., 1998. Variability in dehydrodiferulic acid composition of durum wheat (Triticum durum Desf.) and distribution in milling fractions. Journal of Cereal Science 28, 251258.

Lateral growth of a wheat dough disk under various growth conditions

Amy Penner a, Leaelaf Hailemariam b, Martin Okos a, b, Osvaldo Campanella a, *
a b
School of Agricultural and Biological Engineering, Purdue University, West Lafayette, IN 47907-2093, USA School of Chemical Engineering, Purdue University, West Lafayette, IN 47907-2100, USA
Article history: Received 17 October 2007 Received in revised form 24 June 2008 Accepted 9 July 2008 Keywords: Viscoelasticity Dough Pressure release Expansion
a b s t r a c t
Numerous studies have tried to understand and model bubble growth inside dough. Experimental studies are inconvenienced by the methods inability to capture the dynamic phenomena. In this paper, a versatile experimental method was developed to allow for macroscopic expansion of wheat dough. The study evaluates expansion of a dough disk under varying: moisture content (40, 41, 42, 43, and 44% wb), leavening acid concentration (30, 40, and 50% db), pressure schemas, pressurizing gas (compressed air and CO2), and lubrication (Teon lm coating and Pam aerosol lubricant). Dough expansion increased 22.6% by increasing moisture content from 40 to 44%. Increased baking powder formulation (40% db) was used to enhance initial growth conditions and CO2 production. Pressure pulse and pressure vacuum methods added pressurization alternatively with full vacuum. The former method included a rest period before vacuum application, and increased expansion by 10.8%. Teon and Pam reduced friction between the dough and acrylic plate and increased the nal expansion by 14.7% compared to no lubricant following the standard pressurization method. Pressure pulse and pressure vacuum experiments decreased expansion by 28.4 and 38.2%, respectively compared to standard pressurization while using Teon and Pam. 2008 Elsevier Ltd. All rights reserved.
1. Introduction In the food industry, bubble formation is used to create structure and texture which are primarily responsible for the product attributes (Aguilera and Germain, 2007; Moraru and Kokini, 2003; Trater et al., 2005). Bubbles are generated through the evolution and expansion of gas, either through phase change (ash evaporation), or chemical reaction (yeast or chemical leavening). Gas that does not escape into the atmosphere is bound in nucleating sites which act as growth sites for bubbles in the expanding material (Baker and Mize, 1941). The quantity of bubbles created and sustained is the main factor in determining the nal product quality, texture, and volume (Cauvain et al., 1999; Elmehdi et al., 2003; Fox, 2004; Strybulevych et al., 2007; Whitworth and Alava, 1999). Thus, an understanding of bubble growth mechanics is crucial for control and product optimization. Previous work on mathematical modeling of bubble growth has been done by Alavi et al. (2003), Chen and Rizvi (2006), de Cindio and Correra (1995), Fan et al. (1994, 1999), Schwartzberg et al. (1995), Shimiya and Yano (1988), Singh and Bhattacharya (2005), and Wang et al. (2005). Recently, Hailemariam et al. (2007) presented a model that included the effect of multi-scale mass
* Corresponding author. Tel.: 1 76 5496 6330; fax: 1 76 5496 1115. E-mail address: campa@purdue.edu (O. Campanella). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.07.007
transfer, surface tension, inertia and viscoelasticity in a multibubble environment. However, a major challenge with these models has been the experimental verication of bubble dynamics. Bell et al. (1975) described bubble growth in dough during baking using television microscopy of a thin lm of dough baking on a small slide. Kumagai et al. (1991) used a similar approach, spreading a thin layer of dough on a slide heated by a thermostat. de Cindio and Correra (1995) microscopically analyzed bubble size distributions of dough samples which were immersed in liquid nitrogen and cut in thin slices. Similar analysis was done by Campbell et al. (1998) and Schwartzberg et al. (1995). Alavi et al. (1999) cut extrudate samples into thin sections which were dried in high temperature ovens before electron microscopy analysis. A similar approach was taken by Chen and Rizvi (2006) and Singh and Bhattacharya (2005). With the exception of the dough smear methods of Bell et al. (1975), the experimental methods used thus far have yielded only the nal bubble size distribution, which is of little help in studying bubble dynamics. The dough smear methods do not allow for control of the surrounding environment and are difcult to scale up. Measuring the doughs properties while it is undergoing physical and chemical changes is difcult due to the dynamic and fragile nature of dough (Aguilera and Germain, 2007). Invasion during the expansion process could cause gas cell leakage which would reduce expansion (Wagner et al., 2006). Older techniques used interval
66
A. Penner et al. / Journal of Cereal Science 49 (2009) 6572
measurements, but this yielded inconsistent, time intensive results that were not fully accurate or representative of on-line changes (Fox et al., 2004). Modern microscopic measuring techniques may be applied for dynamic size measurements, but these are not without their own hurdles. These methods include electron microscopy (Aguilera and Germain, 2007; Trater et al. 2005; Whitmore and Alava, 1999), X-ray tomography (Babin et al., 2006, 2007; Bellido et al., 2006; Trater et al., 2005; Whitmore and Alava, 1999), Magnetic resonance imaging (MRI) (Wagner et al., 2006), and ultrasound (Elmhedi et al., 2003; Fox et al., 2004; Strybulevych et al., 2007). The reader is asked to consult these references for greater details. In summary, one may remark that no single experimental method is free of challenges. The ease of experimentation, the size range of the bubbles, and the versatility of the methods should be considered in using or developing an experimental method. The bubble size range in the food industry is 101000 mm (Fyrillas et al., 2000). Current detection methods are applied to determine gas bubble distribution, cell wall size, and cell wall thickness. The microscopic bubble size poses a challenge for the development of a dynamic method of measuring bubble size because the material structure is xed either through heat or freezing before analysis. This may destroy the structure, the entity to be measured, and is not fully representative of the dynamic system (Babin et al., 2007; Trader et al., 2005; Wagner et al., 2006). There is a need for a versatile experimental method that facilitates the dynamic observation of the bubble size history and distribution. To overcome the challenge of measuring bubble growth in situ, the authors developed an experimental method to study microstructure growth through macroscopic investigation. A sample of wheat dough was constrained to grow laterally in a thin gap between parallel plates which allowed the maximum change in density with time due to decreased surface area for gas escape. Some work has been done on lateral growth with a mass of dough, including looking at cookie diameters (Lee and Inglett, 2006; Swanson et al., 1999; Zoulias et al., 2002). Taki et al. (2003) developed an experimental method for analyzing the growth of bubbles in polymers through examination of an expanding disk of uid. Recently, Elmehdi et al. (2007) described a method for measuring dough density between two polyacrylic plates at atmospheric conditions where radial growth was the only dimension encountered. Digital imaging was used to capture dough expansion every 2 min. In this paper, we extend that method and describe a method that allows for dynamic tracking of overall growth under different pressure schema with increased measurement recordings. This work concentrates on the study of rapid dough growth initiated by sudden pressure and vacuum changes. Analysis of the lateral growth in a dough disk may be imagined to be a cylindrical section of the expanding material. It was hypothesized that pressurization during the initial phase of growth would increase the nal expansion due to prevention of early gas escape. To test this idea, different pressurization schemas were implemented to evaluate the effects of pressurization and vacuum application to aid in increasing expansion. The work describes a simple dough system consisting of our, water, and baking powder to observe expansion in experimental testing. Moisture content and baking powder concentration work to drive dough expansion in this system and thus warranted further study to generate optimal testing conditions. Moisture content affects the release of CO2 from the baking powder. It also affects the ability of the dough matrix to expand and retain the gas that is produced. Baking powder concentration was varied to determine its effect on expansion. Use of external CO2 as the pressurizing gas was investigated for its potential effects in the nal system equilibrium. Dough expansion is based on internal CO2 production and thus it was necessary to compare CO2 with compressed air as the
pressurizing gas. It was also hypothesized that expansion was limited by friction at the dough/acrylic plate interface, and thus a reduction in friction was investigated using Teon lm and Pam aerosol lubrication.
2. Experiment description The pressure chamber consists of a clear acrylic cylinder with removable top and bottom clear acrylic plates. The plates sealed the chamber using a rubber gasket at each end and each end was securely bolted to maintain constant pressure once the sample was placed inside. The cylinder measured 10 inches vertically and had a 5 inch inner diameter. A schematic representation of the experimental setup is shown in Fig. 1. A sample of dough was placed between two 5-mm thick clear acrylic plates which were clamped in place with a xed 2 mm gap. This ensured only lateral expansion which is shown in Fig. 2. The acrylic plate was marked with x and y major and minor axes in the horizontal plane using a mm length scale to determine the diameter of dough sample with time. All experiments were performed in triplicate and averaged results are reported. For all experiments temperature was held at 20 3 C. The following parameters were evaluated throughout this work: i. Moisture content: 40, 41, 42, 43, and 44% (wb). ii. Baking powder concentration: 30, 40, and 50% (db). iii. Pressure schema: combinations of increased pressure (2 bars), and/or full vacuum application; see Section 3.2 for full description. iv. Pressurizing gas: compressed air or CO2. v. Lubrication: Teon lm coating and/or aerosol lubrication.
Fig. 1. Expansion chamber: (a) top-schematic; and (b) bottom-laboratory setup.
67
Flour and baking powder were kept in sealed containers in a freezer at 4 C to prevent moisture absorption from the air. They were taken out of the freezer immediately before weighing and stored in the freezer immediately after the sample was measured. Flour and baking powder were blended by stirring 20 times by hand. Water was added all at once and immediately mixed for 2 min at 86 rpm in a 100-g bowl pin mixer (National Manufacturing Company, Lincoln, NE, USA). 3.2. Measurement of lateral expansion Upon completion of mixing, a dough sample (23 g) was placed between the two acrylic plates. Clamps were added to the sides to ensure a constant 2 mm distance between the plates. This also eliminated the vertical expansion component as a variable in experiments. The plates holding the dough were placed inside the pressure chamber and sealed. Periodic photographs show the dough disk expanding as seen in Fig. 3. Measurements were taken of the horizontal major diameters (x and y radial diameters) every minute for 30 min starting 4 min after mixing was completed. The 4 min window was used to mount the sample in the chamber between the clear acrylic plates. Density was determined from the mass of dough (assumed constant) divided by the changing volume of dough.
rt
Where:
m TA
Tp
m
Dh Dv 4
2
Fig. 2. Parallel plate: (a) schematic; and (b) laboratory setup.
3. Materials and methods 3.1. Dough preparation Flour was determined to have 11.5% moisture content (AOAC method 925.10) and baking powders to have 6% moisture content (http://www.nutritiondata.com). The wheat dough was prepared from 10.00 0.02 g all purpose bleached our (Gold Medal General Mills, Minneapolis, MN, USA), 4.00 0.02 g double acting baking powder (Clabber Girl, Terre Haute, IN, USA). A representative formula describes baking powder by mass as starch (40.07%), sodium bicarbonate (26.73%), sodium aluminum sulfate (19.92%) and monocalcium phosphate (13.28%) (Matz, 1992). Room temperature (22 2 C) distilled water was added to produce 40, 41, 42, 43, and 44% moisture percentages on a wet basis. In this work, we used 40% baking powder (db) which is much higher than industrial practices such as 46.25% for baking powder biscuits (Matz, 1992). This was done to accelerate the production of CO2 and enhance expansion in experiments.
A is the area of the dough (mm2) r is density of dough as a function of time m is mass of dough in grams T is sample thickness, constant at 2 mm Dh is the horizontal distance occupied by the dough at each time measurement (mm) Dv is the vertical distance occupied by the dough at each time measurement (mm) This method of computing area was validated using Origin 6.1 Scientic Graphing and Data Analysis Software (OriginLab Corporation, Northampton, MA, USA). Images of the dough disk were digitized, and the area was computed by using numerical integration through Origin 6.1 software. Area measurements from on-line experiments were hand calculated and the results were in close agreement. For example, a hand calculated sample area was 13.50 mm2 while the corresponding software calculated the area at 13.52 mm2. The experiments were monitored, varying: moisture content, baking powder concentration, pressure and vacuum cycles in the
Fig. 3. Change in dough disk dimensions through the pressure scheme: (a) before compression (1 bar); (b) pressurization (2 bar); (c) after release (1 bar). Dotted circle added to show changes in dough dimensions.
68
chamber, pressurizing gas, and the application of lubricants on the acrylic plates. All the experiments, except for those done to study the effect of the pressure prole (discussed later), were done with the standard pressurization method: Minutes 03: The chamber pressure was maintained at 1 bar to allow the dough to rest in the expansion chamber. This allowed for initial CO2 production and temperature equilibration. Minutes 48: Pressure was increased to 2 bars by CO2 or compressed air application to the chamber. Minutes 930: Pressure was immediately released and maintained at 1 bar. Such a pressure prole is necessary because it allows the observer to simulate conditions during rapid expansion or compression of the dough. The diameter was recorded throughout the procedure at 1 min intervals and the dimensionless ratio of density at any time to the initial density was compared. To study pressure effects, the pressure pulse method was examined: Minutes 03: Chamber pressure was maintained at 1 bar to allow the dough to rest in the expansion chamber. Minutes 48: Pressure was increased to 2 bars by CO2 application to the chamber. Minutes 913: Pressure was released and the chamber pressure was maintained at 1 bar. Minutes 1419: A full vacuum was applied to the chamber. Minutes 2030: Vacuum was released and chamber pressure was maintained at 1 bar. The nal pressurization schema used to study pressure effects was the pressure vacuum method: Minutes 03: Chamber pressure was maintained at 1 bar to allow the dough to rest in the expansion chamber. Minutes 48: Pressure was increased to 2 bars by CO2 application to the chamber. Minutes 913: Pressure was released into a full vacuum environment in the chamber and thereby eliminated the re-equilibration phase after pressurization.
Minutes 1430: Vacuum was released and chamber pressure was maintained at 1 bar. 4. Results 4.1. Effect of moisture content (40, 41, 42, 43, and 44%) During experimentation, after pressure was applied the density dramatically increased due to compression of the dough. As moisture content increased from 40 to 44%, greater compression and total expansion were produced. Under the standard pressurization method the high moisture dough (44% wb) yielded a 26% greater expansion compared to the low moisture dough (40% wb) with only 4% expansion as seen in Fig. 4. It was noted that 40 and 41% moisture content doughs were dry and non-cohesive. This was in strong contrast to the 44% dough which was highly viscous. The dough made using 43% moisture content was chosen as the balance between the above extremes and was used in all other experiments. 4.2. Effect of baking powder concentration Formulation consisting of 40% baking powder is signicantly higher than current industrial standards of 4.256% (db) for baking powder products such as biscuits (Matz, 1992). However, this value produced rapid CO2 generation for monitoring dough expansion. Using 6% baking powder (db) in the dough, produced low expansion that also produced poor repeatability when monitoring expansion. The growth effects were seen more easily with 40% (db) baking powder formulation. This concentration was varied to investigate 30, 40, and 50% (db) baking powder in the formula; however there was no signicant difference in the nal expansion. The authors were most concerned with initial conditions, and therefore used 40% baking powder concentration with all reactions because the author had used this with many previous experiments and sought to standardize the concentration. 4.3. Effect of pressure change Lateral expansion was measured under three pressure schemas that were a variation of the standard pressurization method, as
Fig. 4. Change in dough density with varying moisture content where dough followed the standard pressurization method: 1 bar in region A, 2 bar in region B, and 1 bar in region C.
69
described above. The dough was placed directly on the acrylic plates as the effects of friction were explored later. For the atmospheric pressure method, the dough sample was at 1 bar the entire measurement period. This showed the dough disk expanding for 15 min. When the standard pressurization method was applied, the density quickly increased as the dough contracted, and upon release the dough showed gradual expansion. Pressure was introduced into the system to keep the CO2 generated within the compressed dough and demonstrate increased expansion after pressure release. With application of vacuum, the bubble volume was expected to be enhanced. To evaluate the effects of a reduced pressure environment, the standard vacuum method was applied. This followed the standard pressurization method, but instead of applying 2 bars of pressure during minutes 48, a full vacuum was applied. The dough expanded with the vacuum, but upon release, it collapsed and density increased. The use of vacuum produced the least overall expansion with 15.1%. Atmospheric pressure and standard pressurization method gave 17.7, and 19.6% expansion respectively, as represented in Fig. 5. The vertical lines in the gures denote regions of different pressure. 4.3.1. Pressure pulse (with/without vacuum) The pressure pulse experiments were performed by subjecting the dough to large pressure differentials using a combination of 2 bars of pressure or a full vacuum. In all cases, 5 ml Fluoro-Grip-F Optically Clear (HD-FEP) Teon coating (Integument Technologies, Inc., Tonawanda, NY, USA) was applied on the acrylic plates as a lubricant to reduce friction and enhance the effects of multiple changes in pressure. The effect of lubrication is discussed in a later section. The rst and second test case followed the pressure pulse method as described above with the addition of Pam aerosol lubricant (Pam, International Home Foods, Parsippany, NJ, USA) to the second case. The third test case followed the pressure vacuum prole where vacuum was applied immediately after pressurization. The intermediate rest period that occurred during minutes 913 in case 1 and 2 led to a greater expansion than case 3
with 20.7, 11.7, and 1.9%, respectively. Case 3 experienced a continuous drop in pressure which led to a signicantly lower nal expansion. The addition of a pressurization period increased expansion by the potential for distributing gas cells that had begun growth during the initial incubation phase. This is seen in Fig. 6, where vertical lines denote changes in pressure. The repeated change in pressure was performed to study the expansion behavior of the dough after signicant deformation.
4.4. Effect of pressurizing gas (air or CO2) Compressed air and CO2 gas were compared as the pressurizing media to determine if the added environmental CO-2 affected the chemical reaction which produced CO2. Each gas was used to pressurize the chamber to 2 bars following the standard pressurization 4.4.1. Method Using CO2 for pressurization produced a 5.8% greater increase in the nal dough density as seen in Fig. 7. The explanation for this enhancement is that CO2 has a higher solubility in the dough than air. CO2 is produced internally from the baking powder (sodium bicarbonate component) releasing the CO2 from the chemical reaction. Environmental CO2 must be solubilized and transferred to the gaseous phase in the existing gas cells before it can aid in expansion (Eliasson and Larson, 1993). This presents the possibility of diffusion effects when multiple gases, e.g. air and CO2 are present.
Fig. 5. The effect of sudden pressure release and sudden increase on dough expansion where dough followed the standard pressurization method: 1 bar in region A, 2 bar in region B, and 1 bar in region C; atmospheric pressure: 1 bar in region A, B and C; standard vacuum method: 1 bar in region A, full vacuum in region B, and 1 bar in region C.
Fig. 6. The effect of different pressure schema on dough expansion through the pressure pulse method: 1 bar in region A, 2 bar in region B, 1 bar in region C, full vacuum in region D, and 1 bar in region E; and pressure vacuum method: 1 bar in region A, 2 bar in region B, full vacuum in region C, 1 bar in region D and E.
70
Fig. 7. Effect of CO2 versus air as the pressurizing agent on dough expansion where dough followed the standard pressurization method: 1 bar in region A, 2 bar in region B, and 1 bar in region C.
lubrication at 4.9%. Following the pressure release, the greatest expansion was with Teon and Pam which expanded 40.1% compared to Pam at 30.2% or Teon at 20.9%. It is interesting to note that no lubrication (adding neither Teon nor Pam) gave 25.4% expansion which is greater than Teon alone as shown in Fig. 8. However, the difference in using Teon and no lubrication was not signicant. It is of worth to note that in Fig. 6, Teon alone was compared to both Teon and Pam using the pressure pulse method. The combination of Teon and Pam led to an overall expansion of 11.7% compared to Teon alone which was at 20.7%. The author would suggest that the friction between the dough and Teon surface restricted the doughs movement after the full vacuum was released. The effect of lubrication was an indirect study of the viscosity of the dough through the interaction with the stationary plates. The combination of dough and Pam can be considered a two-liquid ow, while the Teon is considered stationary. The difference between the observed behaviors in the two cases would translate into the effect of wall roughness versus a liquid lm. During expansion, bubbles were seen emerging at the Pamdough interface. Monitoring 1 s intervals at the interface, allows one to notice the emergence of bubbles that become entrapped within the Pam liquid. This progression of bubble growth is highlighted with arrows in Fig. 9 which shows the same bubble increasing. The gas released from the dough matrix is believed to be the driving force for creation of the bubbles along the plates. 5. Discussion
4.5. Effect of lubrication The authors were concerned that lateral expansion was inhibited by friction between the acrylic plate and the dough. Reducing this friction using lubricants (Teon and Pam) was done to evaluate the role of friction in expansion. The following four cases were studied with the standard pressurization method: expansion without lubrication, with Teon, with Pam and with both. As expected, during the pressure increase to 2 bars the dough disk compressed the greatest with both Teon and Pam at 30.9%, followed by Pam with 19.0%, Teon with 9.9%, and without
Fig. 8. Effect of lubrication on dough expansion where dough followed the standard pressurization method: 1 bar in region A, 2 bar in region B, and 1 bar in region C.
A series of experiments was performed to study the lateral expansion of dough during rapid expansion and compression. Experiments under varying moisture contents (40, 41, 42, 43, and 44%) showed a decrease in density with increase in the moisture content of the dough. This is consistent with the work of Elmhedi et al. (2007) who showed that dough density decreased as a function of fermentation time for yeast dough using digital imaging. Lai et al. (1989) reported that the addition of sodium bicarbonate contributed a weakened structure in wheat dough. One can observe the dramatic difference in the nal expansion through the change in the pressure schema. It is expected that CO2 is removed more quickly using a full vacuum which reduced the amount of retained gas which would have potentially formed bubbles. The presence of a short residence period of 1 bar pressure following pressurization in the pressure pulse experiments falls in line with this hypothesis: the rapid shift to vacuum would help deprive the dough of CO2. Application of vacuum after an initial growth period was used to understand dough behavior in a reduced pressure environment. The viscoelastic properties of dough may account for the expansion after removal of the deformation induced by pressure or vacuum. The elasticity of the dough is related to the release of potential energy stored during the deformation by either extension or compression forces (Eliasson and Larsson, 1993). The growth phenomena under vacuum and pressure applications appear to qualitatively mirror each other. A comparison of pressurization by compressed air and CO2 showed very similar behavior, with CO2 showing slightly more expansion and compression tendencies. These effects were minor and would not indicate a signicant mass transfer. One possible occurrence for the small discrepancy could be the ability of increased CO2 in the pressurized atmosphere to be incorporated into the dough matrix. The authors were concerned about potential reduction in the doughs nal expansion due to friction at the dough-plate interface. To overcome this challenge, Elmhedi et al. (2007) applied mineral oil during expansion testing to reduce air bubble entrapment at the interface of the measuring vessel and the dough. Teon lm
71
combination of Teon coating with Pam lubrication was compared to only Teon coating as seen in Fig. 6. As expected, combining Teon and Pam yielded the greatest expansion, but it was interesting to note that Teon had little effect by itself. 6. Summary and future work Experimental studies for bubble growth are inconvenienced by the methods inability to capture the dynamic phenomena. In this paper, a versatile experimental method is developed to allow for macroscopic expansion of wheat dough. The study evaluated expansion of a dough disk under varying: moisture content (40, 41, 42, 43, and 44% wb), leavening acid concentration (30, 40, and 50% db), pressure schemas, pressurizing gas composition (compressed air and CO2), and lubrication (Teon lm coating and Pam aerosol lubricant). Dough expansion increased 22.6% by increasing moisture content from 40 to 44% (wb). At 44% moisture (wb), this dough was highly viscous and 43% moisture content was chosen for all experiments. Baking powder was used at an increased level, 40% (db) to enhance initial growth conditions and facilitate rapid CO2 production. Pressure pulse and pressure vacuum methods added pressurization alternatively with full vacuum. The former method included a rest period before vacuum application which increased expansion by 10.8%. The addition of a pressurization period aided expansion through potential for the promotion of distributing gas cells that had begun growth during the initial incubation phase. Teon and Pam reduced friction between the dough and acrylic plate to increase the nal expansion by 14.7% compared to no lubricant (standard pressurization method). Pressure pulse and pressure vacuum experiments decreased expansion by 28.4 and 38.2%, respectively compared to standard pressurization while using Teon and Pam. Several extensions of this work are possible. Work is currently underway to model the reaction rate of CO2 production within a dough matrix. In addition to this, other upcoming extensions are to include the determination of whether the improved expansion for higher moisture content was a result of improved consistency or increased reaction rate. The incorporation of lateral mass transfer could be studied by drilling holes through the acrylic plates. Finally, work is to be done on the study of air entrapment and its effect on lubrication with an emphasis on possible creation of an air cushion between the dough and the acrylic surface. Acknowledgements The authors would like to thank Apri Melinda, Airine Airine, Reinaldo Kusliawan and Louis Nelson III for their invaluable help in developing the methodology. We would also like to thank the SURF program and NASA-SBIR grant for providing the funding. References
Aguilera, J.M., Germain, J.C., 2007. Advances in image analysis for the study of food microstructure. In: McClements, D.J. (Ed.), Understanding and Controlling the Microstructure of Complex Food. Woodhead Publishing Limited, Cambridge, pp. 261287. Alavi, S.H., Gogoi, B.K., Khan, M., Bowman, B.J., Rizvi, S.S.H., 1999. Structural properties of protein-stabilized starch-based supercritical uid extrudates. Food Research International 32, 107118. Alavi, S.H., Rizvi, S.S.H., Harriot, P., 2003. Process dynamics of starch based microcellular foams produced by supercritical uid extrusion I: model development and II. Numerical simulation and experimental evaluation. Food Research International 6, 309330. AOAC International, 2000. Ofcial Methods of Analysis, sixteenth ed. AOAC International, Gaithersburg, MD. Babin, P., Della Valle, G., Chiron, H., Cloetens, P., Hoszowska, J., Pernot, P., Reguerre, A.L., Salvo, L., Dendievel, R., 2006. Fast X-ray tomography analysis of bubble growth and foam setting during breadmaking. Journal of Cereal Science 43 (3), 393397.
Fig. 9. Rapid growth of air bubbles through entrapment at the dough-Pam interface: (a) top at 9:22.00 min; (b) middle at 9:22.30 min; (c) bottom 9:23.00 min.
coating and Pam aerosol spray were used to reduce friction for these experiments. These additions did not appear to affect the qualitative expansion behavior, but they did enhance the expansion and compression effects observed. The compression was signicantly different for pressure pulse experiments when the
72
A. Penner et al. / Journal of Cereal Science 49 (2009) 6572 Kumagai, H., Lee, B.H., Kumagai, H., Yano, T., 1991. Critical radius and time course of expansion of an isolated bubble in wheat our dough under temperature rise. Agricultural and Biological Chemistry 55 (4), 10811087. Lai, C.S., Guetzlaff, J., Hoseney, R.C., 1989. Role of sodium bicarbonate and trapped air in extrusion. Journal of Cereal Chemistry 66, 6973. Lee, S., Inglett, G.E., 2006. Rheological and physical evaluation of jet-cooked oat bran in low calorie cookies. International Journal of Food Science and Technology 41, 553559. Matz, S.A., 1992. Bakery Technology and Engineering, third ed. Van Nostrand Reinhold, New York. Moraru, C.I., Kokini, J.L., 2003. Nucleation and expansion during extrusion and microwave heating of cereal foods. Comprehensive Reviews in Food Science and Food Safety 2, 120138. Schwartzberg, H.G., Wu, J., Nussinovitch, A., Mugerwa, J., 1995. Modeling deformation and ow during vapor-induced pufng. Journal of Food Engineering 25, 329372. Shimiya, Y., Yano, T., 1988. Rates of shrinkage and growth of air bubbles entrained in wheat our dough. Agricultural and Biological Chemistry 52 (11), 2879 2883. Singh, A.P., Bhattacharya, M., 2005. Development of dynamic modulus and cell opening of dough during baking. Journal of Texture Studies 36, 4467. Strybulevych, A., Leroy, V., Scanlon, M.G., Page, J.H., 2007. Characterizing a model food gel containing bubbles and solid inclusions using ultrasound. Soft Matter 3 (11), 13881394. Swanson, R.B., Carden, L.A., Parks, S.S., 1999. Effect of a carbohydrate-based fat substitute and emulsifying agents on reduced-fat peanut butter cookies. Journal of Food Quality 22, 1929. Taki, K., Nakayama, T., Yatsuzuka, T., Ohshima, M., 2003. Visual observations of batch and continuous foaming processes. Journal of Cellular Plastics 39, 155169. Trater, A.M., Alavi, S., Rizvi, S.S.H., 2005. Use of non-invasive X-ray microtomography for characterizing microstructure of extruded biopolymer foams. Food Research International 38 (6), 709719. Wagner, M., Zhang, L., Quellec, S., Doursat, C., Flick, D., Trystram, G., Lucas, T., 2006. Role of the crust formation on local expansion during bread baking. In: Campbell, G.M., Pyle, L., Scanlon, M. (Eds.), Proceedings of Bubbles in Food: Novelty, Health and Luxury, September 1014, Windermere (UK). Wang, L., Ganjyal, G.M., Jones, D.D., Weller, C.L., Hanna, M.A., 2005. Modeling of bubble growth dynamics and nonisothermal expansion in starch-based foams during extrusion. Advances in Polymer Technology 24 (1), 2945. Whitworth, M.B., Alava, J.M., 1999. The imaging and measurement of bubbles in bread doughs. In: Campbell, G.M., Webb, C., Pandiella, S.S., Niranjan, K. (Eds.), Bubbles in Foods. Eagan Press, St. Paul, MN, pp. 221231. Zoulias, E., Oreopoulou, V., Kounalaki, E., 2002. Effect of fat and sugar replacement on cookie properties. Journal of the Science of Food and Agriculture 82, 1637 1644.
Babin, P., Della Valle, G., Dendievel, R., Lourdin, D., Salvo, L., 2007. X-ray tomography study of the cellular structure of extruded starches and its relation with expansion phenomenon and foam mechanical properties. Carbohydrate Polymers 68 (2), 329340. Baker, J.C., Mize, M.D., 1941. The origin of the gas cell in bread dough. Cereal Chemistry 18, 1934. Bell, A.V., Berger, K.G., Russo, J.V., White, G.W., Weathers, T.L., 1975. A study of the micro-baking of sponges and cakes using cine and television microscopy. Journal of Food Technology 10, 147156. Bellido, G.G., Scanlon, M.G., Page, J.H., Hallgrimsson, B., 2006. The bubble size distribution in wheat our dough. Food Research International 39 (10), 10581066. Campbell, G.M., Reilly, C.D., Fryer, P.J., Sadd, P.A., 1998. Aeration of bread dough during mixing: effect of mixing dough at reduced temperature. Cereal Food World 433, 163167. Cauvain, S.P., Whitworth, M.B., Alava, J.M., 1999. The evolution of bubble structure in bread doughs and its effect on bread structure. In: Campbell, G.M., Webb, C., Pandiella, S.S., Niranjan, K. (Eds.), Bubbles in Foods. Eagan Press, St. Paul, MN, pp. 8588. Chen, K.H., Rizvi, S.S.H., 2006. Rheology and expansion of starch-water-CO2 mixtures with controlled gelatinization by supercritical uid extrusion. International Journal of Food Properties 9, 863876. de Cindio, B., Correra, S., 1995. Mathematical modeling of leavened cereal goods. Journal of Food Engineering 24, 379403. Eliasson, A.C., Larson, K., 1993. Cereals in Breadmaking A Molecular Colloidal Approach. Marcel Dekker, Inc, New York. Elmehdi, H.M., Page, J.H., Scanlon, M.G., 2003. Using ultrasound to investigate the cellular structure of bread crumb. Journal of Cereal Science 38, 3342. Elmehdi, H.M., Page, J.H., Scanlon, M.G., 2007. Evaluating dough density changes during fermentation by different techniques. Journal of Cereal Chemistry 84, 2007. Fan, J., Mitchell, J.R., Blanchard, J., 1994. A computer simulation of the dynamics of bubble growth and shrinkage during extrudate expansion. Journal of Food Engineering 23, 337356. Fan, J., Mitchell, J.R., Blanchard, J.M.V., 1999. A model for the oven rise of dough during baking. Journal of Food Engineering 41, 6977. Fox, P., Smith, P.P., Sahi, S., 2004. Ultrasound measurements to monitor the specic gravity of food batters. Journal of Food Engineering 65 (3), 317324. Fyrillas, M.M., Kloek, W., van Vliet, T., Mellema, J., 2000. Factors determining the stability of a gas cell in an elastic medium. Langmuir 16, 10141019. Hailemariam, L., Okos, M., Campanella, O., 2007. A mathematical model for the isothermal growth of bubbles in wheat dough. Journal of Food Engineering 82, 466477. http://www.nutritiondata.com/facts-C00001-01c21Ke.html Retrieved on 8/28/ 2007.

Digestibility of protein and starch from sorghum (Sorghum bicolor) is linked to biochemical and structural features of grain endosperm
Joshua H. Wong a, Tsang Lau a, Nick Cai a, Jaswinder Singh a, 2, Jeffrey F. Pedersen b, William H. Vensel c, William J. Hurkman c, Jeff D. Wilson d, Peggy G. Lemaux a, *, Bob B. Buchanan a,1
a
Department of Plant and Microbial Biology, University of California, Berkeley, CA 94720, USA United States Department of Agriculture, Agricultural Research Service, Lincoln, NE 68583, USA United States Department of Agriculture, Agricultural Research Service, Western Regional Research Center, Albany, CA 94710, USA d United States Department of Agriculture, Agricultural Research Service, Manhattan, KS 66502, USA
b c
Article history: Received 21 December 2007 Received in revised form 9 July 2008 Accepted 9 July 2008 Keywords: Disulde proteins Starchprotein interface Granule-bound starch synthase I Sorghum
a b s t r a c t
Although a principal source of energy and protein for millions of the worlds poorest people, the nutritional value of sorghum (Sorghum bicolor L. Moench) is diminished because of low digestibility of grain protein and starch. To address this problem, we analyzed the properties of two sorghum lines that have a common pedigree but differ in digestibility. Consistent with results based on a ruminal uid assay, the protein and starch of one line (KS48) was more thoroughly digested than that of the other (KS51) using in vitro assays based on pepsin and a-amylase. The indigestibility of KS51 relative to KS48 was shown to be due to (i) a greater abundance of disulde-bonded proteins; (ii) presence in KS51 of nonwaxy starch and the accompanying granule-bound starch synthase; and (iii) the differing nature of the protein matrix and its interaction with starch. The current ndings suggest that each of these factors should be considered in efforts to enhance the nutritional value of sorghum grain. 2008 Elsevier Ltd. All rights reserved.
1. Introduction Grain sorghum (Sorghum bicolor L. Moench) ranks fth in worldwide production among cereal crops, after wheat, rice, corn and barley. The U.S. is the number two producer and the number one exporter of sorghum, primarily to Mexico for use as animal feed (U.S. Grains Council, 2006). Popularity of sorghum is due in part to its ability to produce reasonable yields in warmer, drier regions. Because subsistence farmers in Africa and Asia cultivate sorghum widely as a staple food for home consumption, the crop is
Abbreviations: DTT, dithiothreitol; GBSS, granule-bound starch synthase; IVDMD, in vitro dry matter disappearance; ME, 2-mercaptoethanol; Mr, relative molecular mass; PAGE, polyacrylamide gel electrophoresis; SDS, sodium dodecyl sulfate; MOPS, 3-(N-morpholino) propane sulfonic acid. * Corresponding author. Tel.: 1 5106421589; fax: 1 5106427356. E-mail addresses: viewmont@nature.berkeley.edu (J.H. Wong), shout_at_neko@ hotmail.com (T. Lau), monto@nature.berkeley.edu (N. Cai), jaswinder.singh@ mcgill.ca (J. Singh), jeff.pedersen@ars.usda.gov (J.F. Pedersen), vensel@pw.usda.gov (W.H. Vensel), bhurkman@pw.usda.gov (W.J. Hurkman), jeff.d.wilson@ars.usda.gov (J.D. Wilson), lemauxpg@nature.berkeley.edu (P.G. Lemaux), view@nature. berkeley.edu (B.B. Buchanan). 1 Submitting author. 2 Present address: McGill University, Macdonald Campus, 21,111 Lakeshore Road, Ste-Anne-de-Bellevue, Quebec H9X 3V9, Canada. 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.07.013
a principal source of energy and protein for millions of the worlds poorest (Klopfenstein and Hoseney, 1995). Seed proteins of sorghum are less digestible than those of other cereals and digestibility is exacerbated by wet cooking the meal or our, which results in signicant nutritional losses (Hamaker et al., 1986; MacLean et al., 1981; Mertz et al., 1984). Numerous factors contribute to the digestibility problem [for review, see (Duodu et al., 2003)]. Exogenous aspects include the interaction of protein with non-protein components, such as polyphenols, starch, nonstarch polysaccharides, phytates and lipids. Endogenous factors arise from the nature of the proteins themselves and their organization within the grain (Belton et al., 2006; Duodu et al., 2003; Ezeogu et al., 2005, 2008; Hamaker and Bugusu, 2003). As in other cereals, the low amount of protein relative to starch, i.e., approximately 10% protein, vs. 7080% starch, based on grain dry weight (Rooney and Pugfelder, 1986), affects the functional properties of starch, such as gelatinization and digestion rate, to a greater extent in sorghum than in other cereals (Chandrashekar and Kirleis, 1988; Duodu et al., 2002; Ezeogu et al., 2005, 2008). The proteins of the sorghum grain are classically divided, based on solubility in different solvents (Jambunathan et al., 1975; Landry and Moureaux, 1970): albumins (water-soluble), globulins (salt-soluble), karins (prolamins, aqueous alcohol-soluble), cross-linked karins (aqueous alcohol reducing agent-soluble), cross-linked glutelins (detergent reducing agent alkaline pH-soluble) and unextracted
74
J.H. Wong et al. / Journal of Cereal Science 49 (2009) 7382
structural protein residue. A newer and more simplied classication scheme for sorghum proteins has been proposed that divides them into two groups, karins and non-karins. This scheme is based on the homogeneous nature and varied origin of the karin storage prolamins relative to the heterogeneous nature of the non-karin proteins (i.e., albumins, globulins and glutelins) that are involved in cellular functions (Hamaker and Bugusu, 2003; Hamaker et al., 1995). Despite a high degree of sequence homology to maize zeins, sorghum storage proteins contain a higher proportion of cross-linked fractions and are more hydrophobic, explaining their greater propensity to form intermolecular disulde-cross linkages and possibly additional protein aggregates that could facilitate the formation of more covalent bonds compared to maize zeins (Belton et al., 2006; Hamaker and Bugusu, 2003). Karins, comprising 7080% of the protein in whole grain sorghum our (Hamaker et al., 1995), are synthesized and translocated into the lumen of the endoplasmic reticulum where they form protein bodies (Taylor et al., 1985). They are subclassied as: a- (23 and 25 kDa), b- (20 kDa) and g- (28 kDa) types based on molecular weight, extractability, structure and cross-reactivity with sera against analogous maize zeins (Belton et al., 2006; Mazhar et al., 1993; Shull et al., 1991). Comprising w80% of total karins, the a-type is considered the principal storage protein, followed by the g- (w15%) and b- (w5%) members. Recently, a DNA-derived sequence for a d-karin was shown to have high homology with the Mr 10,000 d-zein, except for absence of part of the methionine-rich region (Belton et al., 2006; Izquierdo and Godwin, 2005). a-Karins have low levels of cysteine relative to the b- and g-types (5 and 7 mol%, respectively), whereas d-karins are rich in methionine (1618 mol%) but lack cysteine. Electron microscopy of internal protein body structure reveals that g-karins, and to a lesser extent b-karins, encapsulate the more digestible a-karins in a disuldebond polymer network (Oria et al., 2000; Shull et al., 1992), thereby impeding exposure to proteases. In the corneous endosperm, non-karins (albumins, globulins, glutelins) form around protein bodies, effectively gluing the bodies into a matrix surrounding the starch granules (Hamaker and Bugusu, 2003; Shull et al., 1990; Taylor et al., 1984). This protein matrix appears to act as a barrier to starch gelatinization and digestibility (Chandrashekar and Kirleis, 1988; Duodu et al., 2002; Ezeogu et al., 2005, 2008) due to cross-linking between g- and bkarins and matrix proteins (Duodu et al., 2001; Hamaker and Bugusu, 2003; Oria et al., 1995a). Cooking reduces digestibility by effecting a conformational change in proteins that could facilitate formation of disulde-linked polymers (Axtell et al., 1981; Duodu et al., 2002, 2003; Hamaker et al., 1987; Oria et al., 1995b). The negative impact of cooking on protein digestibility was mitigated by addition of 2-mercaptoethanol (ME) or other reducing agents (Elkhalifa et al., 1999; Hamaker et al., 1987). Sorghum grains rich in karin-containing protein bodies also have a lower capacity for starch gelatinization (Chandrashekar and Kirleis, 1988; Ezeogu et al., 2005, 2008)dan observation consistent with the nding that adding ME during cooking to cleave disulde bonds within the protein matrix increased the degree of starch gelatinization and digestion (Elkhalifa et al., 1999; Ezeogu et al., 2005, 2008; Zhang and Hamaker, 1998). The protein barrier surrounding the starch granule may also reduce the hydrolysis of native and processed starch by amylolytic enzymes (Rooney and Pugfelder, 1986). Thus, the addition of pronase to hydrolyze the protein matrix signicantly enhanced in vitro rates of starch hydrolysis by increasing surface area and enabling starch to interact with a-amylase and amyloglucosidase (Rooney and Pugfelder, 1986). Similar effects were observed when sorghum our was either treated with pepsin before cooking or cooked in the presence of dithiothreitol (DTT) or other reductants (Zhang and Hamaker, 1998). Abundance of starch granules may also
decrease proteolysis by limiting accessibility of proteolytic enzymes, especially when gelatinized during cooking (Duodu et al., 2002). Uniqueness of the protein matrix and its interaction with starch that affect the rate of starch digestion are key differences between the feed quality of sorghum and corn (Rooney and Miller, 1982; Rooney and Pugfelder, 1986). In addition, endosperm texture and cooking conditions have been shown to have a signicant effect on in vitro digestibility of starch and protein in these two cereals (Duodu et al., 2002, 2003; Ezeogu et al., 2005, 2008). Sorghum displays signicant variation in rates of starch disappearance (Wester et al., 1992). Based on a 12-h incubation in the in vitro dry matter disappearance (IVDMD) assay, loss of dry matter in sorghum correlated closely with rates of starch, but not necessarily protein digestion (Pedersen et al., 2000). Collectively, these ndings suggest that properties of starch and protein in sorghum could affect their mutual digestibility. The majority of published references deal with digestibility of the proteins and their impact on starch, but evidence for the reverse is scant (Duodu et al., 2002). To pursue this issue and assess the importance of relevant factors in a single study, we investigated how protein and starch inuence the breakdown of one another by comparing two sorghum lines having a common pedigree but differing in digestibility.
2. Materials and methods 2.1. Grain and preparation of materials The sorghum lines KS48 and KS51 were selected for these experiments because of previously determined differences in digestibility but the identical pedigree, Texioca-63 Short Kaura (Casady, 1972). Both have a clear pericarp and yellow endosperm. Field-grown seeds were produced at the University of Nebraska Agricultural Research Development Center, Ithaca, NE; additional seeds were generated from greenhouse-grown plants at the University of California, Berkeley. Seed size and hardness of KS48 and KS51 were assessed using a Perten SKCS 4100 Single Kernel Characterization System instrument. The starch content of KS48 and KS51 was 64.7% and 67.8%, respectively, on a dry weight basis. For 12-h IVDMD analysis, mature dry seeds were ground in a Wiley mill to pass a 20-mesh screen. For all other analyses, seeds were ground in a Wiley mill to pass a 40-mesh screen. Ground meal was stored at 25 C until used in digestion assays. Pepsin (porcine stomach mucosa, P-7000) and three types of a-amylases bacterial (A-3403, Type XII-A), porcine pancreas (A-3176, Type VI-B) and human saliva (A-1031, Type XIII-A) were used in digestibility assays (Sigma, St Louis, MO).
2.2. Microscopy studies For microscopic observations, seeds were sectioned with a razor blade and stained with iodine (0.3 g I2, 1.0 g KI in 100 ml water). Sectioned seeds were used for electron microscopic analyses, using a Hitachi TM 1000 environmental scanning electron microscope (Hitachi High Technologies America, Inc., Pleasanton, CA). Prior to analysis, sections were coated with gold-palladium to 1 mm using a Tousimis sputter machine (Rockville, MD).
2.3. Twelve-hour IVDMD of sorghum meal Sorghum meal was digested for 12 h in rumen uid inoculum obtained from a ruminally stulated steer using previously described methods (Pedersen et al., 2000).
75
2.4. Sequential protein extraction Sorghum meal (2 g) was sequentially extracted with the six solvents listed below in 20 ml lots at 25 C and centrifuged (18,900g, 10 min) at 4 C after each step [modied from (Hamaker et al., 1986)]. Initially, meal was extracted with 0.5 M NaCl by shaking for 60 min. Resulting NaCl-soluble albumins and globulins (fraction 1) were retained, and the pellet sequentially extracted yielding the following supernates: double distilled H2O (ddH2O) for 20 min (water-wash fraction, fraction 2); 60% 2-propanol (v/v) for 4 h (karins, fraction 3); 0.1 M borate buffer, pH 10.8, for 4 h (glutelin-like protein, fraction 4); 60% 2-propanol with 1% dithiothreitol (DTT) for 4 h (cross-linked karins, fraction 5); and 0.1 M borate buffer, pH 10.8, containing 1% DTT and 1% sodium dodecyl sulfate (SDS) for 18 h at 4 C (cross-linked glutelins, fraction 6). Protein fractions were stored at 20 C. 2.5. Disulde protein content of selected fractions Equal amounts of protein in fractions extracted without reducing agent were reduced with 1 mM DTT; protein sulfhydryl groups were labeled with monobromobimane (mBBr) and analyzed using SDS-PAGE (Wong et al., 2004). Relative intensity of uorescence of labeled proteins was equated to relative amount of protein disuldes. 2.6. Rapid in vitro pepsin digestion
boiled 5 min, centrifuged and run in 12% BisTris gels with MOPS buffer for 1 h 20 min at 150 V. Fluorescent imaging and protein staining were as described (Section 2.11). 2.11. Quantication of uorescent proteins To determine redox status, uorescent (unsaturated and saturated) images of reduced proteins were captured from gels on a Gel Doc 1000 imager using Quantity One software (Bio-Rad, Hercules, CA) with 365 nm UV light either immediately or after a 15 min water-wash; images were saved in TIFF format. Gels were stained for protein with colloidal Coomassie Blue G-250 overnight at 25 C, destained in ddH2O and scanned with a UMAX PowerLook 1100 scanner using UMAX MagicScan version 4.4 within Adobe Photoshop, v.6 (Adobe Systems, San Jose, CA). Amount of protein, expressed as volume (intensity area) of uorescent bands and/or undigested karin bands, was quantied with Quantity One software. Higher protein band volume indicates more protein in the gel, and thus lower digestibility (Aboubacar et al., 2001). 2.12. 2-D gel electrophoresis Protein contained in equal volumes of extracts was precipitated by addition of 5 vol of cold (20 C) acetone and incubation at 20 C overnight. Proteins were recovered by centrifugation and separated by 2-D gel electrophoresis (Wong et al., 2004). 2.13. Western blot analysis
Time course of in vitro pepsin digestion of sorghum meal and processing of the resulting undigested residue were as described (Aboubacar et al., 2001; Nunes et al., 2004). 2.7. Extraction of protein undigested by pepsin Protein in pellets, not digested by pepsin, was extracted with 0.0125 M borate buffer, pH 10 containing 1% SDS and 2% 2-ME as described (Aboubacar et al., 2001). The extraction was modied further to include extraction of a granule-bound starch synthase I (GBSSI) from the borateSDSME extracted residue (see Section 2.8). 2.8. Extraction of granule-bound starch synthase
After electrophoresis, proteins were transferred to nitrocellulose membranes at 4 C for 75 min at 50 V constant voltage (Wong et al., 2004) and cross-reactivity analyzed with waxy protein antibody (Terada et al., 2000). 2.14. In vitro starch digestion Digestibility of starch in uncooked sorghum meal was measured using a modied hydrolysis method (Zhang and Hamaker, 1998) with bacterial a-amylase (A-3403, Type XII-A, Sigma Chemical Co., St. Louis, MO). 2.15. Sequential addition of a-amylase/pepsin or pepsin/a-amylase
To determine if unextracted protein remained in residues after extraction with borateSDSME (0.0125 M borate, 1% SDS, 2% ME, pH 10), the washed pellets (in 0.5 ml of the same buffer) were mixed with 1 ml borateSDSME buffer and extracted for 510 min in a boiling water bath and cooled in tap water. Gelatinized starch was pelleted by centrifugation (20,800g, 10 min) and the clear supernate was saved for SDS-PAGE analysis. 2.9. SDS-PAGE A Criterion Pre-Cast Gel System using TrisHCl gels (Bio-Rad, Hercules, CA) was used for SDS-PAGE analysis (Wong et al., 2004). Equal amounts of protein were used for protein distribution analyses and estimation of sulfhydryl content. For protein digestion analysis, equal volumes (w5 ml) with a specied amount of extracted protein were used. 2.10. SDS-PAGE of karins and glutelins undigested by pepsin NuPAGE Novex BisTris gels (Invitrogen, Carlsbad, CA), with a neutral pH to minimize protein modications, were used to obtain optimal resolution of small- to medium-sized proteins. Aliquots (w5 ml) of specic amounts of extracted protein were mixed with 1/4 volume of 4 NuPAGE sample buffer plus 0.4% ME,
2.15.1. a-Amylase followed by pepsin Meal (50 mg) was treated with 50 mg a-amylase (human saliva form, Sigma A-1031, Type XIII-A) in 0.3 ml 1 mM Na glycerophosphateHCl buffer (25 mM NaCl, 5 mM CaCl2, 0.02% w/v azide), pH 6.9, at 37 C for 1 h with shaking. Pepsin solution (1 ml at 20 mg/ ml, pH 2.0) was added, the mixture incubated for 2 h at 37 C and neutralized with 0.1 ml 2 N NaOH. Supernate containing a-amylaseand pepsin-digested material was collected by centrifugation (20,800g, 10 min) for reducing sugar determination. Pellet was washed with 1 ml 0.1 M KH2PO4 buffer, pH 7.0 and then 1 ml ddH2O and extracted for remaining karins and GBSSI as above. Controls were: (A) a-amylase alone, (B) pepsin alone and (C) buffer alone. 2.15.2. Pepsin followed by a-amylase Meal (50 mg) was incubated with pepsin (1 ml at 20 mg/ml, pH 2.0) for 2 h at 37 C with shaking and then neutralized (0.1 ml 2 N NaOH to pH 7.0). a-Amylase (50 mg) in 0.3 ml Na glycerophosphate buffer as above, was added and the mixture incubated for 1 h at 37 C. Reaction was stopped by acidication and the supernate containing pepsin- and a-amylase-digestible materials was saved for reducing sugar determination. Pellets were washed 2 with 1 ml ddH2O and extracted for karins and GBSSI (Sections 2.7 and 2.8). Controls were: (A) pepsin alone, (B) a-amylase alone and (C) buffer alone.
76
2.16. Effect of DTT reduction on digestion of cross-linked karins and glutelins by pepsin Meal (50 mg) was incubated 5 mM DTT in 0.5 ml of 0.1 M KH2PO4, pH 7.0, for 1 h at 37 C and supernate removed by centrifugation (20,800g, 10 min). The DTT-treated pellet was washed with 1 ml (A) ddH2O and (B) 0.1 M KH2PO4 buffer, pH 2.0. The washed residue was incubated with pepsin (1 ml at 20 mg/ml in pH 2.0 KH2PO4 buffer) at 37 C for 2 h with shaking and digestion was stopped with 0.1 ml 2 N NaOH. Pepsin-treated residue was extracted for remaining karins and glutelins and analyzed by NuPAGE as above.
3.1. Protein distribution of KS48 and KS51 The distribution of protein in fractions extracted with the indicated solvents (Section 2.4) suggested that, in general, the two sorghum lines yielded approximately the same amount of total extractable protein (Table 1). However, differences in distribution of protein between the two lines suggested that the more digestible KS48 contains more protein than the less digestible KS51 in the rst ve fractions extracteddddH2O wash, albumin and globulin (saltsoluble), karin (60% 2-propanol), glutelin-like protein (borate soluble), and cross-linked karin (2-propanol DTT). By contrast, the least soluble (last-extracted) fraction, cross-linked glutelin (borate DTT SDS), represented a considerably greater fraction in KS51 than in KS48. To determine whether these differences in protein distribution affect digestibility, we examined the composition and redox status of protein in the six fractions (Fig. 1). Disulde bonds between cysteine residues, a common feature of most cereal seed proteins, have long been known to stabilize various molecular structures (Wall, 1971). Many proteins with disulde bonds resist digestion by proteases (Astwood et al., 1996; Opstvedt et al., 1984; del Val et al., 1999). Since a large percentage of sorghum karin storage proteins exist in polymeric forms linked by disulde bonds in their native state (Duodu et al., 2002, 2003; El Nour et al., 1998; Hamaker et al., 1987; Oria et al., 1995b), differences in content of fractions rich in insoluble disulde proteins, i.e., 2-propanol DTT (cross-linked karin) and borate DTT SDS (cross-linked glutelin) could contribute to protein digestibility differences. Accordingly, although KS48 had more 2-propanol DTT extractable proteins (cross-linked karins), the content of borate DTT SDS-extractable glutelins was appreciably greater in KS51 than in KS48 a reection of the relative abundance of intermolecular SS groups in this fraction (Table 1). The latter could be a factor in the lower digestibility of KS51. 3.2. Estimation of disulde protein status Protein disulde content was estimated by determining the sulfhydryl groups resulting from DTT reduction, using a thiolspecic uorescent probe, mBBr (Wong et al., 2004). Four fractions (nos. 14 in Table 1), extracted without reducing agent, were monitored for DTT-induced sulfhydryl changes. KS51 showed more intense uorescence relative to KS48 in fraction 1 salt (albumin and globulin) and fraction 4 borate (glutelin-like) (Fig. 1B). The more pronounced increase in proteins with disulde bonds in fraction 6 (borate DTT SDS) could also contribute to the lower digestibility of KS51, as could the markedly different distribution of proteins in that fraction (Fig. 1A). For reasons not yet understood, the uorescence intensity of the karin (propanol) fraction showed larger relative differences between the two lines in untreated vs. DTT-treated samples (Fig. 1B, lanes 9 vs. 11; 10 vs. 12). In summary,
2.17. Granule-bound starch synthase (GBSSI) identication Coomassie blue-stained spots were excised from 2-D gels and placed in a 96-well reaction tray. After proteins were destained, reduced, alkylated and cleaved with trypsin by a robotic sample handler, they were analyzed by tandem mass spectrometry and identied as described (Balmer et al., 2006).
2.18. Other analytical procedures Protein was estimated in buffer-soluble extracts by the dyebinding method and those in the presence of SDS and ME or DTT by the Non-Interfering Protein Assay (Geno Technology, Inc., St. Louis, MO) as described (Wong et al., 2004). The ratio of amylose and amylopectin was determined using a Megazyme Amylose/ Amylopectin Assay Kit (Megazyme, Bray, Ireland).
3. Results and discussion The 12-h IVDMD assay estimates the in vitro rate of starch digestion, which is highly correlated to feed efciency or gain/feed ratio (r 0.94) (Stock et al., 1987). Because it mimics biochemical conditions in cattle, use of ruminal uid containing hydrolytic enzymes produced by the ruminant and its microbial ora is relevant to studies of this type. The 12-h IVDMD applied to the two sorghum lines, KS48 and KS51, gave digestibility values of 33.0% and 22.1%, respectively. The results revealed a difference in dry matter disappearance and, hence, in the rates of digestion of starch as well as protein. The difference revealed by IVDMD, plus the common pedigree of these two lines, led to their choice for further study of digestibility. In the present work, grain from the two lines are subjected to in vitro pepsin and a-amylase assays that are generally considered to reect activities of the human stomach and small intestine, respectively (Astwood et al., 1996; Zhang and Hamaker, 1998).
Table 1 Protein distribution of two sorghum lines with a common pedigree differing in digestibility KS48a Fraction 1 2 3 4 5 6 Extraction solvent NaCl, 0.5 M H2O-wash 2-Propanol, 60% Borate, 0.1 M pH 10.8 2-Propanol, 60% 1% DTT Borate, 0.1 M pH 10.8 1% DTT 1% SDS Total mg protein extracted 28.3 0.5 0.4 1.0 17.3 1.0 5.7 0.7 30.9 2.8 66.3 3.1 % Total protein extracted 19.0 0.3 11.6 3.8 20.8 44.5 KS51a Total mg protein extracted 27.6 0.4 0.4 1.4 12.4 1.3 4.4 1.2 27.8 4.8 73.8 4.3 % Total protein extracted 18.8 0.3 8.5 3.0 19.0 50.4 KS48KS51
D, %
0.2 0.0 3.1 0.8 1.8 5.9
Total
a
148.9 1.5
100
146.4 2.2
100
Mean of 5 experiments.
77
Fig. 1. Effect of DTT on redox status of protein fractions extracted from grain of KS48 and KS51. (A): Protein stained with Coomassie blue; (B): protein labeled with mBBr visualized by UV absorption. TrisHCl gel was used. Individual protein fractions are identied above (A).
the relative abundance of intermolecular SS groups in the crosslinked glutelin fraction in KS51 suggests that the non-karins, which make up the protein matrix, have a strong inuence on the digestibility of KS51 relative to KS48. 3.3. Time course of in vitro pepsin digestion The gel-based system used to analyze protein digestion by pepsin (Aboubacar et al., 2001; Nunes et al., 2004) requires less material, is faster and can accommodate more samples than a commonly used batch procedure (Mertz et al., 1984). The gel procedure not only reveals the types of undigested karins, but also
allows an estimate of the percentage digested over time. This gel system has been successfully used to separate sorghum proteins (Bean, 2003) and wheat glutenins (Kasarda et al., 1998). In our study, we observed that separation of different karin types is improved in the neutral pH environment of the NuPAGE Novex Bis Tris gels relative to that seen with the conventional TrisHCl gels (cf. Fig. 2A vs. Fig. 1A). A NuPAGE gel image (Fig. 2A) shows that high Mr components of the insoluble protein fraction, including glutelins (35100 kDa), were readily digested, while the smaller karins (Mr 1827 kDa) were more resistant. Further, the time course of the in vitro digestion of this fraction demonstrated that karins in KS48 were digested more rapidly by pepsin than those in KS51 (Fig. 2A).
Fig. 2. Time course of in vitro digestion of storage protein and starch in KS48 and KS51. (A): BT-NuPAGE gel showing digestion of glutelins and karins over time; (B): estimation of glutelin and karin digestion rate with the NI protein assay. Results were calculated by linear regression and expressed as % of control without pepsin; (C): starch digestion by bacterial a-amylase. Corn starch is shown as a control.
78
Gel densitometry measurements revealed that various karins, a, b and g (El Nour et al., 1998), were digested at different rates: b > g > a (data not shown). Furthermore, data on the disappearance with time of insoluble proteins (karins and glutelins) (Fig. 2B) indicated that KS48 was digested more rapidly than KS51 negative slope of 0.643 (SE 0.010) for KS48 vs. 0.530 (SE 0.023) for KS51, mean of three determinations. These values, which represent a 22% rate difference, positively correlate with the corresponding IVDMD difference of 36%. The lower value obtained with the in vitro pepsin method could be due to use of a single protease vs. a mixture of proteolytic and amylolytic enzymes present in ruminal uid used in the IVDMD procedure. 3.4. Time course of in vitro starch digestion by a-amylase The nding that the 12-h IVDMD value was higher for KS48 than KS51 suggested that starch digestion alone might follow the same trend, and this was conrmed (Fig. 2C). The average of three timecourse experiments for in vitro digestion of meal from KS48 and KS51 by bacterial a-amylase gave slopes of 1.74 103 (SE 1.69 104) and 1.41 103 (SE 5.01 105), respectively. These results represent a 23% difference in starch digestion rate between the two lines a value similar to that obtained with the in vitro pepsin (Fig. 2B) and IVDMD (Pedersen et al., 2000) procedures. More signicantly, the three independent in vitro techniques used to assess digestibility showed a positive correlation. This is the rst time the three approaches have been compared in a single study. Ezeogu et al. (2005) made signicant contributions to the idea that in vitro digestibility of starch in sorghum and maize ours is related to endosperm texture and cooking conditions. These researchers found that starch digestion was signicantly higher in oury sorghum endosperm than in vitreous endosperm, while their counterparts in maize were similar in this regard. Cooking with ME increased starch digestion in both sorghum and maize, but more so with sorghum, and with vitreous endosperm ours. This is consistent with the fact that more polymeric karins, formed by intermolecular disulde bonds (Chandrashekar and Mazhar, 1999; Oria et al., 1995a), occurred in the vitreous endosperm fraction (Ezeogu et al., 2005; Kumari and Chandrashekar, 1994), probably because the cysteine-rich g- and b-karins abound in that part of the sorghum grain (Chandrashekar and Mazhar, 1999; Ezeogu et al., 2005, 2008; Mazhar and Chandrashekar, 1995). Unlike their ndings, which were based on similar types of starch i.e., a 75:25 amylopectin/amylose ratio from different parts of the sorghum and maize endosperm our results are based on different types of
starch, i.e., an amylopectin/amylose ratio of 97:3 and 70:30, for KS48 and KS51, respectively (see below).
3.5. Starch digestibility differences in starch in KS48 and KS51 In the original rapid protein digestion assay (Aboubacar et al., 2001), the pellet remaining after extraction of karins and glutelins with borateSDSME was not further analyzed. In this study, we attempted to ll this gap by analyzing the residual pellet that still contained between 2.22.5% (pepsin-treated) and 4.75.0% (buffer control) of protein nitrogen (data not shown). After boiling the residual pellet with excess borateSDSME buffer, we unexpectedly observed a marked difference in the gelatinization properties of starch in KS48 and KS51. The KS48 pellet yielded a transparent, soft starch gel that swelled to a greater extent than the KS51 counterpart. Further, when equal amounts of extraction buffer were added to the pellets before boiling, only a small amount of supernate was recovered after centrifuging the cooled gel KS48 preparation, whereas KS51 yielded a larger amount of clear supernate that separated easily from the harder opaque gel (data not shown). The gelatinization differences in KS48 and KS51 starch resemble those observed for waxy vs. non-waxy starch. In addition to obvious swelling differences in gelatinized starch, supernates from KS48 and KS51 contained different proteins. Onedimensional SDS-PAGE analysis revealed a major protein band of Mr w 58 kDa in KS51, not detected in KS48 (Fig. 3A). Because it was released from starch granules only after gelatinization, the protein did not appear to change after treating the parent meal with pepsin (data not shown). Two-dimensional gel analysis gave information on the nature of the unknown band (Fig. 3C, D). Two major 61 kDa spots with acidic pIs, uniquely observed in KS51 (note arrows, Fig. 3D), were identied as granule-bound starch synthase I by mass spectrometry [GBSSI, chloroplast (sorghum) accession No. Q43134]. Identity was further conrmed by Western blot analysis using an antibody against maize waxy protein (Fig. 3B). In KS51, a highly antibody-reactive band was seen at 58 kDa and a less reactive one at 61 kDa, corresponding to two isoforms of GBSSI: a major form in the endosperm (58 kDa) and a minor form in the pericarp (61 kDa) (Nakamura et al., 1998). While lower levels of the pericarp isoform were seen in KS48 relative to KS51, the endosperm component was not detected. The granule-bound isoform of starch synthase, GBSSI, also referred to as the waxy protein encoded in the waxy (Wx) locus of cereals, functions specically to elongate amylose (Denyer et al., 2001) by adding glucose residues in a-1,4-glycosidic
Fig. 3. Starch-bound proteins extracted from KS48 and KS51. The pellet remaining after removing glutelins and karins was boiled with borateSDS2-ME buffer and the proteins solubilized were analyzed by one-dimensional SDS-PAGE. (A): Protein; (B): Western blot analysis using waxy protein antibody; (C and D): two-dimensional gel analysis of starchbound proteins extracted in (A).
79
linkage. Recent genetic experiments have revealed that sorghum contains at least two GBSS genes (Pedersen et al., 2005). 3.6. Starch of waxy and non-waxy sorghum Normal or non-waxy sorghum starch has approximately 75% amylopectin and 25% amylose, while waxy starch contains nearly 100% amylopectin (Ezeogu et al., 2005; Rooney and Miller, 1982; Rooney and Pugfelder, 1986). The apparent absence of GBSSI protein in the endosperm of KS48 (Fig. 3A, B) indicates that its starch should be low in amylose and contain mainly amylopectin. Waxy and non-waxy polymers of starch can be distinguished by iodine staining (Denyer et al., 2001; Pedersen et al., 2004). Starch of non-waxy lines, containing amylose and amylopectin, forms blueblack complexes with iodine, while starch from waxy lines, lacking amylose, stains reddish brown. Since the waxy nature of KS48 and KS51 was not indicated in the USDA-ARS, Genetic Resource Information Network (Pedersen et al., 2004), grains from the two lines were stained with iodine. Results conrmed KS48 has a waxy (brownish red) and KS51 a non-waxy (dark blue) type of endosperm (see below). Biochemical analyses conrmed this observation: KS48 contains 3% and KS51 30% amylose. Numerous studies have shown that starch is more readily digested in waxy sorghum than in non-waxy types [for review, see (Rooney and Pugfelder, 1986)]. Our results are consistent with this conclusion, not only for starch but also for protein (KS48 > KS51). Digestibility of protein, however, may not necessarily be consistent with waxy vs. non-waxy sorghum lines. The nature of the protein matrix and the extent of embedded starch in the endosperm are proposed to account for this inconsistency (Chandrashekar and Kirleis, 1988; Duodu et al., 2003; Rooney and Miller, 1982; Rooney and Pugfelder, 1986; Shull et al., 1990). To gain insight into this point, we conducted experiments with our waxy and non-waxy lines as described below. 3.7. Sequential addition of a-amylase and pepsin Digestion protocols, using sequential addition of pepsin followed by a-amylase, and vice versa, were designed to study the inter-relationship of starch and protein in KS48 and KS51. Prior digestion with pepsin, followed by treatment with human saliva aamylase (designated P/A in Table 2), enhanced to a greater extent starch hydrolysis, relative to treatment with a-amylase followed by pepsin (A/P), in KS48 compared to KS51 (1.8- vs. 1.5-fold, respectively). This observation suggests that KS51 starch granules were intrinsically harder to digest than those of KS48, even if the protein matrix was partially removed by prior pepsin digestion. This explanation agrees with the time course of in vitro starch digestion of meal (Fig. 2C), as well as with the relative ease of digestion of isolated waxy vs. non-waxy starch (Rooney and Miller, 1982; Rooney and Pugfelder, 1986). Reversal of this protocoldremoval of starch prior to pepsin digestiondshowed the opposite effect; proteolysis was enhanced to a greater extent with KS51 than KS48 (2.2- vs. 1.6-fold) (Table 3).
Table 3 Protein digestion by pepsin without and with prior removal of starch by a-amylase Sorghum line With a-amylase pre-treatment (a-amylase/pepsina) Protein remaining, volume, % KS48 KS51
a
Without a-amylase pre-treatment (pepsin/a-amylasea) Protein remaining, volume, % 40.9 2.7 34.0 3.8
Ratio A/P:P/A
65.4 2.0 75.3 1.5
1.6 2.2
This unexpected result suggests that, in addition to the intrinsic properties of the starch, spatial arrangement between the granules and protein bodies differs in the two lines. Further, starch in KS51 appears not only to be harder to digest with a-amylase, but also more capable of impeding protein digestion by pepsin than in KS48. Once starch-imposed restrictions are removed, protein in KS51 appears more exposed and thus more accessible to pepsin. A question arises as to whether the observed difference could simply be due to the physical properties of the two grains, i.e., variation in the hardness of the two grains, or different ratios of hard (closed) to soft (open) endosperm, as reported by Ezeogu et al., 2005. Hardness data for KS48 and KS51 samples using the single kernel characterization system (SKCS) showed that they are practically and statistically the same, with a mean for seed hardness of 69.12 and 69.83, respectively. Therefore, our interpretation is that the difference observed above is mostly due to the properties of the starch and proteins. 3.8. Effect of DTT reduction on digestion of karins and glutelins by pepsin Reducing agents, such as 2-ME (Elkhalifa et al., 1999; Ezeogu et al., 2005, 2008; Hamaker et al., 1987) and DTT or thioredoxin (del Val et al., 1999), were shown to improve protein digestibility, thus highlighting the importance of components with disulde bonds (Hamaker et al., 1987). In assessing DTT effects, we observed that 42% and 50% of the karins of KS48 and KS51, respectively, were not digested by pepsin in control samples. Reduction with DTT
Table 2 Starch digestion with and without prior removal of protein by pepsin Sorghum line With pepsin pre-treatment (pepsin/a-amylasea) mg reducing sugar/h KS48 KS51
a
Without pepsin pre-treatment (a-amylase/pepsina) mg reducing sugar/h 47.6 2.1 39.3 1.3
Ratio P/A:A/P
84.8 4.2 59.1 2.7
1.8 1.5
Fig. 4. Effect of redox status on digestion of the combined glutelin and karin fractions of KS48 and KS51 by pepsin. D P DTT pepsin; P pepsin; D DTT; C control (minus DTT and pepsin). * Karin region.
80
enhanced digestibility by 3.5-fold and 2.0-fold, with undigested protein decreasing from 42% to 12% for KS48 and from 50% to 25% for KS51 (Fig. 4, * designates karin region). To achieve a level of karin digestion with KS51 equivalent to that of KS48, either a longer incubation period or more reductant was required (data not shown). 3.9. Microscopy studies Our results thus far show that a greater abundance of disulde proteins and the non-waxy type of starch contribute to the low digestibility of KS51 vs. KS48. We have performed experiments with additional waxy and non-waxy lines of sorghum and corn (3 pairs of sorghum lines, 1 pair of corn lines). Digestion of waxy starch was consistently faster than that of non-waxy starch in all samples tested but protein digestibility was mixed (data not shown). Thus, this data also show the importance of the interactions between starch and protein to the digestibility of both
components. However, they provide no information on the nature of the differences observed in these interactions between the two lines. To study this point, we examined iodine-stained sections of grains microscopically and found that the endosperm in KS51 was more highly organized than that of KS48di.e., there was a distinct demarcation between oury and corneous endosperm in KS51, not present in KS48 (Fig. 5A). Taken together, the analytical data presented above and the iodine staining results conrmed the earlier classication of the lines as waxy (KS48) and non-waxy (KS51) (Pedersen et al., 2004, 2005). Electron micrographs revealed organizational differences between the two. While oury endosperm was similar (Fig. 5C), corneous endosperm differed markedly in the two lines (Fig. 5B). Protein bodies were more numerous and more tightly associated with starch granules in KS51 than in KS48 (Fig. 5B). This type of interaction between the two components could contribute to the low digestibility of protein and starch in KS51.
Fig. 5. Light and electron microscopy of KS48 and KS51 seeds. A. Cross section of seeds stained with iodine showing oury (F) and corneous (C) endosperm. Magnication: 32. B. Electron micrograph of corneous endosperm of KS48 and KS51, 5000. C. Electron micrograph of oury endosperm of KS48 and KS51, 5000. SG starch granule; PB protein body; M protein matrix.
81
In summary, our results highlight the importance of disulde proteins and GBSSI in conferring the chemical and structural properties that inuence the digestibility of stored reserves, both protein and starch. These ndings also add insight into the role of the protein matrix and its interaction with starch. Comparison of two sorghum lines with a common pedigree that differed in digestibility made it possible to demonstrate how proteinstarch interactions in the seed inuence the digestibility of each. Attempts to improve the digestibility of grain by classical breeding and genetic engineering should further our understanding of these factors and their relative contribution to digestibility. These efforts may also provide answers to related fundamental questions, such as what role the key redox protein, thioredoxin, plays in forming and maintaining the protein matrix (Buchanan and Balmer, 2005) and the importance of this unique packaging of stored reserves to the structure and physiology of seed. Answers to these questions will in all likelihood be relevant to human and animal nutrition and to the use of sorghum as a biofuel. Acknowledgements This study was supported by the U.S. Agency for International Development and the Bill and Melinda Gates Foundation. PGL was supported by the USDA Cooperative Extension Service and BBB by the Agricultural Experiment Station through UC Berkeley. We thank Drs. K. McDonald and G. Min (Electron Microscope Laboratory) and D. Schichnes (Biological Imaging Laboratory) for assistance with the microscopy studies and Dr. S.R. Wessler for the gift of waxy protein antibody. The late Dr. K. Kobrehel helped launch this study during a sabbatical visit to Berkeley in the early 1990s. Numerous undergraduate students also made contributions, especially D. Bergmann, P.-H. Ren and P. Yu-Man Kei. References
Aboubacar, A., Axtell, J.D., Huang, C.-P., Hamaker, B.R., 2001. A rapid protein digestibility assay for identifying highly digestible sorghum lines. Cereal Chemistry 78. Astwood, J.D., Leach, J.N., Fuchs, R.L., 1996. Stability of food allergens to digestion in vitro. Nature Biotechnology 14, 12691273. Axtell, J.D., Kirleis, A., Hassen, M.M., DCroz-Mason, N., Mertz, E.T., Munck, L., 1981. Digestibility of sorghum proteins. Proceedings of the National Academy of Sciences of the United States of America 78, 13331335. Balmer, Y., Vensel, W.H., DuPont, F.M., Buchanan, B.B., Hurkman, W.J., 2006. Proteome of amyloplasts isolated from developing wheat endosperm presents evidence of broad metabolic capability. Journal of Experimental Botany 57, 15911602. Bean, S.R., 2003. Evaluation of novel pre-cast gradient gels for the separation of sorghum proteins. Cereal Chemistry 80, 500504. Belton, P.S., Delgadillo, I., Halford, N.G., Shewry, P.R., 2006. Karin structure and functionality. Journal of Cereal Science 44, 272286. Buchanan, B.B., Balmer, Y., 2005. Redox regulation: a broadening horizon. Annual Reviews of Plant Biology 56, 187220. Casady, A.J., 1972. 10 Sorghum parental lines. Crop Science 12, 722. Chandrashekar, A., Kirleis, A.W., 1988. Inuence of protein on starch gelatinization in sorghum. Cereal Chemistry 65, 457462. Chandrashekar, A., Mazhar, H., 1999. The biochemical basis and implications of grain strength in sorghum and maize. Journal of Cereal Science 30, 193207. Denyer, K., Johnson, P., Zeeman, S., Smith, A.M., 2001. The control of amylose synthesis. Journal of Plant Physiology 158, 479487. Duodu, K.G., Nunes, A., Delgadillo, I., Parker, M.L., Mills, E.N.C., Belton, P.S., Taylor, J.R.N., 2002. Effect of grain structure and cooking on sorghum and maize in vitro protein digestibility. Journal of Cereal Science 35, 161174. Duodu, K.G., Tang, H., Grant, A., Wellner, N., Belton, P.S., Taylor, J.R.N., 2001. FTIR and solid state 13C spectroscopy of proteins of wet cooked and popped sorghum and maize. Journal of Cereal Science 33, 261269. Duodu, K.G., Taylor, J.R.N., Belton, P.S., Hamaker, B.R., 2003. Factors affecting sorghum protein digestibility. Journal of Cereal Science 38, 117131. El Nour, I.N.A., Peruffo, A.D.B., Curioni, A., 1998. Characterisation of sorghum karins in relation to their cross-linking behaviour. Journal of Cereal Science 28, 197207. Elkhalifa, A.E.O., Chandrashekar, A., Mohamed, B.E., El Tinay, A.H., 1999. Effect of reducing agents on the in vitro protein and starch digestibilities of cooked sorghum. Food Chemistry 66, 323326.
Ezeogu, L.I., Duodu, K.G., Taylor, J.R.N., 2005. Effects of endosperm texture and cooking conditions on the in vitro starch digestibility of sorghum and maize ours. Journal of Cereal Science 42, 3344. Ezeogu, L.I., Duodu, K.G., Emmanbux, M.N., Taylor, J.R.N., 2008. Inuence of cooking conditions on the protein matrix of sorghum and maize endosperm ours. Cereal Chemistry 85, 397402. Hamaker, B.R., Bugusu, B.A., 2003. Overview: sorghum proteins and food quality. Paper presented at: Workshop on the proteins of sorghum and millets: enhancing nutritional and functional properties for Africa [CD] (Pretoria: South Africa). Hamaker, B.R., Kirleis, A.W., Butler, L.G., Axtell, J.D., Mertz, E.T., 1987. Improving the in vitro protein digestibility of sorghum with reducing agents. Proceedings of the National Academy of Sciences of the United States of America 84, 626628. Hamaker, B.R., Kirleis, A.W., Mertz, E.T., Axtell, J.D., 1986. Effect of cooking on the protein proles and in vitro digestibility of sorghum and maize. Journal of Agricultural and Food Chemistry 34, 647. Hamaker, B.R., Mohamed, A.A., Habben, J.E., Huang, C.-P., Larkins, B.A., 1995. An efcient procedure for extracting maize and sorghum kernel proteins reveals higher prolamin contents than the conventional method. Cereal Chemistry 72, 583588. Izquierdo, L., Godwin, I.D., 2005. Molecular characterization of novel methioninerich delta-karin seed storage protein gene in sorghum (Sorghum bicolor L.). Cereal Chemistry 82, 706. Jambunathan, R., Mertz, E.T., Axtell, J.D., 1975. Fractionation of soluble proteins of high-lysine and normal sorghum grain. Cereal Chemistry 52, 119121. Kasarda, D.D., Woodard, K.M., Elva, A.A., 1998. Resolution of high molecular weight glutenin subunits by a new SDS-PAGE system incorporating a neutral pH buffer. Cereal Chemistry 75, 7071. Klopfenstein, C.F., Hoseney, R.C., 1995. Nutritional properties of sorghum and the millets. In: Dendy, D.A.V. (Ed.), Sorghum and Millets: Chemistry and Technology. American Association of Cereal Chemists, St. Paul, MN, pp. 125168. Kumari, S.R., Chandrashekar, A., 1994. Relationships between grain hardness and the contents of prolamins and three anti-fungal proteins in sorghum. Journal of Cereal Science 20, 9399. Landry, J., Moureaux, T., 1970. Heterogeneite des gluenines du grain de mais: Extraction slelective et composition en acides amines des trios fractions isolees. Bulletin de la Societe de Chimie Bioologique 52, 10211031. MacLean, W.C., de Romana, G.L., Placko, R.P., Graham, G.G., 1981. Protein quality and digestibility of sorghum in preschool children: balance studies and plasma free amino acids. Journal of Nutrition 111, 19281936. Mazhar, H., Chandrashekar, A., 1995. Quantitation and distribution of karins in the kernels of sorghum cultivars varying in endosperm vitreousness. Journal of Cereal Science 21, 155162. Mazhar, H., Chandrashekar, A., Shetty, H.S., 1993. Isolation and immunochemical characterization of alcohol-extractable proteins (karins) of Sorghum bicolor (L.) Moench. Journal of Cereal Science 17, 8393. Mertz, E.T., Hassen, M.M., Cairn-Whittern, C., Kirleis, A.W., Tu, L., Axtell, J.D., 1984. Pepsin digestibility of proteins in sorghum and other major cereals. Proceedings of the National Academy of Sciences of the United States of America 81, 12. Nakamura, T., Vrinten, P., Hayakawa, K., Ikeda, J., 1998. Characterization of a granule-bound starch synthase isoform found in pericarp of wheat. Plant Physiology 118, 451459. Nunes, A., Correia, I., Barros, A., Delgadillo, I., 2004. Sequential in vitro pepsin digestion of uncooked and cooked sorghum and maize samples. Journal of Agricultural and Food Chemistry 52, 20522058. Opstvedt, J., Miller, R., Hardy, R.W., Spinelli, J., 1984. Heat-induced changes in sulfhydryl groups and disulde bonds in sh protein and their effect on protein and amino acid digestibility in rainbow trout (Salmo gairdneri). Journal of Agricultural and Food Chemistry 32, 924935. Oria, M.P., Hamaker, B.R., Axtell, J.D., Huang, C.-P., 2000. A highly digestible sorghum mutant cultivar exhibits a unique folded structure of endosperm protein bodies. Proceedings of the National Academy of Sciences of the United States of America 97, 50655070. Oria, M.P., Hamaker, B.R., Shull, J.M., 1995a. In vitro protein digestibility of developing and mature sorghum grain in relation to a-, b- and g-karin disulde crosslinking. Journal of Cereal Science 22, 8593. Oria, M.P., Hamaker, B.R., Shull, J.M., 1995b. Resistance of sorghum a-, b- and gkarins to pepsin digestion. Journal of Agricultural and Food Chemistry 43, 21482152. Pedersen, J.F., Bean, S.R., Funnell, D.L., Graybosch, R.A., 2004. Rapid iodine staining techniques for identifying the waxy phenotype in sorghum grain and waxy genotype in sorghum pollen. Crop Science 44, 764767. Pedersen, J.F., Bean, S.R., Graybosch, R.A., Park, S.H., Tilley, M., 2005. Characterization of waxy grain sorghum lines in relation to granule-bound starch synthase. Euphytica 144, 151156. Pedersen, J.F., Milton, T., Mass, R.A., 2000. A twelve-hour in vitro procedure for sorghum grain feed quality assessment. Crop Science 40, 204208. Rooney, L.W., Miller, F.R., 1982. Variation in the structure and kernel characteristics of sorghum. Paper presented at: International Symposium on Sorghum Grain Quality (ICRISAT Center, Patancheru, India). Rooney, L.W., Pugfelder, R.L., 1986. Factors affecting starch digestibility with special emphasis on sorghum and corn. Journal of Animal Science 63, 16071623. Shull, J.M., Chandrashekar, A., Kirleis, A.W., Ejeta, G., 1990. Development of sorghum [Sorghum bicolor (L.) Moench] endosperm in varieties of varying hardness. Food Structure 9, 253267. Shull, J.M., Watterson, J.J., Kirleis, A.W., 1991. Proposed nomenclature for the alcohol-soluble proteins (karins) of Sorghum bicolor (L. Moench) based on
82
J.H. Wong et al. / Journal of Cereal Science 49 (2009) 7382 U.S. Grains Council, 2006. Sorghum, grain sorghum: any of various plants of the genus sorghum family Poaceae, a cereal grain. del Val, G., Yee, B.C., Lozano, R.M., Buchanan, B.B., Ermel, R.W., Lee, Y.M., Frick, O.L., 1999. Thioredoxin treatment increases digestibility and lowers allergenicity of milk. Journal of Allergy and Clinical Immunology 1103, 690697. Wall, J.S., 1971. Disulde bonds: determination, location and inuence on molecular properties of proteins. Journal of Agricultural and Food Chemistry 19, 619. Wester, T.J., Gramlich, S.M., Britton, R.A., Stock, R.A., 1992. Effect of grain sorghum hybrid on in vitro rate of starch disappearance and nishing performance of ruminants. Journal of Animal Science 70, 28662876. Wong, J.H., Cai, N., Tanaka, C.K., Vensel, W.H., Hurkman, W.J., Buchanan, B.B., 2004. Thioredoxin reduction alters the solubility of proteins of wheat starchy endosperm: an early event in cereal germination. Plant and Cell Physiology 45, 407415. Zhang, G., Hamaker, B.R., 1998. Low a-amylase starch digestibility of cooked sorghum ours and the effect of protein. Cereal Chemistry 75, 710713.
molecular weight, solubility, and structure. Journal of Agricultural and Food Chemistry 39, 8387. Shull, J.M., Watterson, J.J., Kirleis, A.W., 1992. Purication and immunocytochemical localization of karins in Sorghum bicolor (L. Moench) endosperm. Protoplasma 171, 6474. Stock, R.A., Brink, D.R., Kreikemeier, K.K., Smith, K.K., 1987. Evaluation of earlyharvested and reconstituted grain sorghum in nishing steers. Journal of Animal Science 65, 548556. Taylor, J.R.N., Novellie, L., Liebenberg, N.V.D.W., 1984. Sorghum protein body composition and ultrastructure. Cereal Chemistry 61, 69. Taylor, J.R.N., Schussler, L., Liebenberg, N.V.D.W., 1985. Protein body formation in the starchy endosperm of developing Sorghum bicolor (L.) Moench seeds. S-Afr Tydskr Plantk 51, 8186. Terada, R., Nakajima, M., Isshiki, M., Okaagaki, R.J., Wessler, S.R., Shimamoto, K., 2000. Antisense waxy genes with highly active promoters effectively suppress waxy gene expression in transgenic rice. Plant and Cell Physiology 41, 881888.

Fundamental study on protein changes taking place during malting of oats

Christina Klose, Beatus D. Schehl, Elke K. Arendt*
Department of Food and Nutritional Sciences, National University of Ireland, University College Cork, College Road, Cork, Ireland
Article history: Received 28 February 2008 Received in revised form 8 July 2008 Accepted 9 July 2008 Keywords: Oats Malting Proteins Lab-on-a-Chip analysis
a b s t r a c t
During the malting process, storage proteins are degraded by proteolytic enzymes into small peptides and amino acids. The activity of these enzymes was measured during malting of oats and was found to be increased. To quantify proteolytic degradation, proteins of unmalted, germinating and malted grains were fractionated. After extracting the oat proteins (Osborne fractionation), protein fractions were analysed using a Lab-on-a-Chip technique, which separates the proteins based on their molecular weight by capillary electrophoresis. This new technique for the analysis of proteins was supported by using two-dimensional gel electrophoresis. In addition, amino acid analysis was carried out. In general a degradation of proteins to small peptides and amino acids could be observed in the globulin, prolamin and glutelin fractions. In the albumin fraction a protein increase was observed, which is due to the fact that this fraction contains the majority of the metabolically active proteins. Amino acid analysis supported the observation of increased protein amount in the albumin fraction and decreased protein amounts in the other fractions. Some proteins, which have not been described in the literature, were detected in the albumin and glutelin fraction, since Lab-on-a-Chip technique allows detection of proteins with low molecular weights of 4.5 kDa. 2008 Elsevier Ltd. All rights reserved.
1. Introduction Oats (Avena sativa) is one of the most popular cereals for human consumption and they have received increased interest because of their excellent health-related properties, such as high contents of dietary bre, especially b-glucan, as well as minerals and antioxidants. The most abundant antioxidants are vitamin E (tocols), phytic acid, phenolic compounds and avenanthramides; these are concentrated in the outer layers of the kernel (Peterson, 2001). The major component of the endosperm cell walls of the oat grain or caryopsis is a (1 / 3), (1 / 4)-b-D-glucan, known as oat b-glucan (Ren et al., 2003). The early interest in oat b-glucan arose from knowledge of its highly viscous properties and therefore potential commercial value as a thickening agent in food formulations and other industrial applications, which is interesting for healthy gluten-free products. Whether oats is gluten-free is not yet clear. However, oats can be tolerated by most but not all people with coeliac disease and in Finland 69% of patients were reported to use oat products (Peraaho et al., 2004). For this purpose, a gluten-free beer out of malted oats could also be attractive for coeliacs. Malted oats has been widely used as an ingredient for beer production in medieval times and before. Nowadays oat malt is used in the brewing industry mainly as a avour adjunct for the
* Corresponding author. Tel.: 353 21 490 2064; fax: 353 21 427 0213. E-mail address: e.arendt@ucc.ie (E.K. Arendt). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.07.014
production of special lagers, ales and stouts. Malting is the initial step in beer production and strongly denes type and quality of the beer. The main purpose of malting is to produce enzymes and to breakdown cell walls surrounding starch granules. One of the most important physicalchemical changes that occur during malting is the degradation of the proteinaceous matrix that surrounds the starch granules within the cells of the endosperm and their conversion into soluble peptides and amino acids to provide substrates for the synthesis of proteins in the growing embryo (Briggs et al., 1981). Malt proteins have a high impact on the brewing process and the subsequent quality of beer. Protein content and size distribution are of particular interest in terms of ltering, fermentability, foam and haze stability. Compared to the commonly used grain for malting, barley and oats has not only a unique protein composition, but also a high protein content of 1115 %, of which about a third passes into the nal beer. A number of authors (Lapvetelainen et al., 1995; Ma and Harwalkar, 1984; Robbins et al., 1971; Robert et al., 1985, 1983a; Shewry, 1995; Wu, 1983; Zarkadas, 1982) have reported on the protein composition of oats. In general cereal proteins have been classied based on their solubility (Osborne fractionation), albumins (water soluble), globulins (salt water soluble), prolamins (soluble in dilute alcohol solutions) and glutelins (soluble in acids or bases). The difference between oats and other cereal grains in the structure of the proteinaceous components is consistent with the differences in the distribution of the protein fractions. Oats lacks the protein matrix, which is characteristic of wheat and some other cereals. In wheat
84
C. Klose et al. / Journal of Cereal Science 49 (2009) 8391
and some other cereals, the storage proteins are insoluble in salt solutions. In terms of oats, a large portion of the salt water-soluble globulins also belong to the storage proteins of the endosperm. Although the literary data varies widely, oats contains a relatively low quantity of prolamins and a high amount of globulins, up to 80% of the total oat protein (Lasztity, 1996). The prolamins (avenins) of oats account for about 15% of the total protein and have, similar to other cereal prolamins, mainly a storage function (Robert et al., 1983b). Glutelins have typically been difcult to completely solubilise and therefore values from 5 to 66% have been reported, depending on the extraction solvent and concentration (Robert et al., 1985). Most of the metabolically active proteins of oats are in the water-soluble albumin fraction. Nevertheless, it cannot be excluded that the globulin fraction and probably the glutelin fraction contain this type of protein. Oat albumins account for 112% of the total protein (Lasztity, 1996). Oat protein distribution affects also the amino acid composition of oats. The generally higher lysine content of albumins and globulins causes the relatively high lysine content in oats, compared to other cereals. In addition, glutamic acid and proline contents are relatively lower (Lasztity, 1996). The aim of this study was to evaluate the changes in protein and amino acid composition from the raw oats through germination to the nal malt. For this purpose, a novel analysis technique, the Labon-a-Chip capillary electrophoresis, was used and was supported by using two-dimensional gel electrophoresis. This is the rst report on a detailed analysis of the protein changes during malting of oats. 2. Experimental 2.1. Materials The oat grain was harvested in 2005 and obtained from Raisio, Finland. Malting was carried out using a micro malting machine (Joe White Malting Systems, Perth, Australia). For this purpose, 8 kg of oat seeds were steeped three times, each stage containing a wet stage at 13 C (6 h, 4 h, 3 h) and an air rest stage at 18 C (10 h, 7 h, 1 h). After a total of 31 h of steeping, seeds were germinated for 5 days and 6 h at three stages (17 C, 15 C and 13 C) and kilned in six stages for 23 h using a schedule that started at 55 C and increased to 85 C. 2.2. Total protein analysis of oats Samples were taken daily throughout the malting process. Preliminary analysis of all samples with the Lab-on-a-Chip system revealed degradation of all proteins during the process, whereas largest changes could be observed during the germination process and only minor changes after germination (before kilning). Thus, unmalted, 3-day-germinating and kilned (malted) samples were taken and freeze-dried (VirTus Benchtop 4k, Biopharma Process Systems, Winchester, UK). Roots were removed and samples were milled with a Buhler Miag laboratory-scale disc mill (Buhler GmbH, Braunschweig, Germany) set at a ne setting of 0.05 mm and stored at 80 C. The proteolytic enzyme activity during the malting process was measured according to the method of Brijs et al. (2002), where haemoglobin was used as substrate. After incubation, the reaction was stopped by the addition of 10% (w/v) trichloroacetic acid. The free a-amino nitrogen levels of the supernatants were assayed with trinitrobenzene-sulfonic acid reagent (0.3%, v/v, in 0.2 M sodium phosphate buffer, pH 8.0) using L-leucine as standard. For this purpose, the supernatant and trinitrobenzene-sulfonic acid reagent were incubated and reaction was stopped with 0.2 M HCl. The absorbance was measured at 340 nm. One unit of activity corresponds to the enzyme activity that liberated 1 mg of L-leucine/h
under the assay conditions. Total nitrogen contents of the unmalted, germinating and malted samples were determined by combustion (according to Analytica-EBC, method 4.3.2) using a Leco FP-528 nitrogen determinator (St Joseph, MI). 2.3. Modied Osborne fractionation of oat proteins The extraction procedure was carried out once, where rst, unmalted, germinating and malted oats (10 g of each sample) were defatted with 100 mL hexane using a Soxhlet extractor. The defatted samples were then extracted twice with 100 mL of distilled water (albumin fraction). After water extraction the residue was extracted twice with 100 mL of 5.0% NaCl (globulin fraction). The remaining our was then extracted three times with 150 ml of 55% 1-propanol 1% DTT (prolamin fraction) and the glutelin fraction together with the residue was extracted three times with a solvent containing 6 M urea, 2% SDS and 1% DTT. Each extraction was carried out at room temperature for 10 min and centrifuged afterwards at 10,000 rpm for 10 min. All supernatants were collected, further dialyzed against distilled water for 24 h, freeze-dried and stored at 80 C. 2.4. Lab-on-a-Chip analysis of total protein composition To evaluate protein changes during the malting process, a novel analysis technique, the Lab-on-a-Chip capillary electrophoresis, was used. The principles of these electrophoretic assays are based on traditional gel electrophoresis principles that have been transferred to a chip format. The chip accommodates sample wells, gel wells and a well for an external standard (ladder). Micro-channels are fabricated in glass to create interconnected networks among these wells. During chip preparation, the micro-channels are lled with a sieving polymer and uorescence dye. Once the wells and channels are lled, the chip becomes an integrated electrical circuit. Charged biomolecules like proteins are electrophoretically driven by a voltage gradient, similar to slab gel electrophoresis. Because of a constant mass-to-charge ratio and the presence of a sieving polymer matrix, the molecules are separated by size. Smaller fragments are migrating faster than larger ones. Dye molecules intercalate into protein micelles. These complexes are detected by laser-induced uorescence. Data is translated into gel-like images (bands) and electropherograms (peaks). With the help of a ladder that contains components of known sizes, a standard curve of migration time versus fragments size is plotted. From the migration times measured for each fragment in the sample, the size is calculated. Two markers are run with each of the samples bracketing the overall sizing range. The lower and upper markers are internal standards used to align the ladder data with data from the sample wells. This is necessary to compensate for drift effects that may occur during the course of a chip run (Agilent Technologies, 2005). For analysis of the total protein content, 40 mg of milled oat samples were extracted at room temperature for 5 min with 400 mL 2 M urea solution containing glycerol, TrisHCl and 0.1 M DTT and centrifuged for 15 min. Four microliters of each supernatant were denatured using 2 mL of Agilent denaturing solution and heated for 5 min at 100 C. After dilution with deionised water, 6 mL were applied to the Protein 230 LabChip and the Protein 80 LabChip for analysis in the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA). 2.5. Lab-on-a-Chip analysis of extracted protein fractions Ten micrograms of each extracted protein fraction were dissolved in 1 mL of their extraction solvents and applied to the Agilent 2100 Bioanalyzer, as described above. For the 230 LabChip,
85
each run included a ladder comprising reference proteins of 7, 15, 28, 46, 63, 95, 150 kDa plus an upper marker of 240 kDa and a lower marker of 4.5 kDa. Each sample contained an internal standard comprising the upper and lower marker as well. Any peak detected below 14 kDa is termed a system peak and is not included in analysis. The detection performance was also carried out in a molecular weight range between 4.5 and 95 kDa using the Protein 80 LabChip. For this analysis the ladder consisted of reference proteins of 3.5, 6.5, 15, 28, 46, 63 kDa plus the upper and the lower markers of 95 and 1.6 kDa. According to the Agilent manual any peak detected below 5 kDa is named a system peak and is not included in analysis. Analyses were carried out in triplicate on each LabChip and a relative standard deviation of 8% was considered. 2.6. Two-dimensional gel electrophoresis of extracted protein fractions Initially oat protein fractions of unmalted, germinating and malted samples were defrosted. All samples were then dissolved in a solubilisation buffer containing 9 M urea, 4% CHAPS, 0.05% Triton X100 and 65 mM DTT. Protein solution (125 mL) was applied to each strip. Isoelectrofocusing (IEF) was carried out using 7-cm IPG 310 strips (ReadyStrip, BioRad) and a BioRad PROTEAN IEF cell with controlled cell temperature of 20 C. The running conditions were as follows: passive rehydration: 8 h, active rehydration (50 V): 8 h, rapid (300 V): 30 min, linear (4000 V): 20,000 V-h, rapid (300 V): 99 h. After IEF was completed, the IPG strips were equilibrated for 15 min in a buffer containing 50 mM TrisHCl pH 8.8, 6 M urea, 30% glycerol, 2% SDS and130 mM DTT. After that, the strips were equilibrated for another 15 min in the same solution except that DTT was replaced with 130 mM iodoacetamide and traces of bromophenol blue. SDS-PAGE was performed in a BioRad Criterion Dodeca cell with gels of total acrylamide concentration of 12.5% at 20 C, sealed with Tris/glycine/SDS running buffer. The gels were run at 100 V until the tracking dye reached the bottom of the gel. To visualize the proteins, the gels were rst incubated in a xation solution, containing ethanol and phosphoric acid and then in an incubation solution containing methanol, phosphoric acid and ammonium sulphate dissolved in distilled water. The staining was carried out with the incubation solution containing Coomassie brilliant blue for 4 days. Analysis was carried out in triplicate. 2.7. Amino acid analysis The free amino acid prole of the milled samples was measured in triplicate using a Jeol JLC-500/V amino acid analyser (Jeol (UK) Ltd., Garden City, Herts, UK) tted with a Jeol Na high performance cation exchange column. Prior to analysis, samples were extracted with a 12% (w/v) trichloracetic acid solution. This procedure was carried out three times. For the total amino acid prole, milled samples as well as each protein fraction were analysed three times according to the method of Moore and Stein (1963). Due to this method, tryptophan, which is not stable during hydrolysis, could not be identied, and glutamine and asparagine were converted to their acids during hydrolysis, so that they were detected with their acid. 3. Results 3.1. Proteolytic activity Total proteolytic enzyme activity was analysed using haemoglobin as substrate and absorption was measured against L-leucine as standard. During the malting process an 8.6-fold increase in proteolytic activity could be observed. The results revealed an increase in the proteolytic activity level also during steeping (from
2.4 to 4.5 mg of L-leucine/h/g), during 3 days of germination (from 4.5 to 16.3 mg of leucine/h/g) and during the last steps of malting (from 16.3 to 20.3 mg of leucine/h/g). The highest increase could be detected during the rst 3 days of germination. 3.2. Total protein analysis of oats The results of total protein analysis using the Protein 230 LabChip showed several protein peaks in the area of 2330 kDa (area 1), as well as in the area of 3852 kDa (area 2) in all three samples (results not shown). Peak heights in area 1 demonstrate that these proteins were degraded by 47%, from about 500 FU (Fluorescence Units) to about 200 FU during malting. In area 2, peak heights decreased by 74% after malting. Identical protein patterns were obtained with the Protein 80 LabChip (results not shown). Both peak areas decreased by 44% (1) or by 71% (2) during the malting process. However, the Protein 80 LabChip reveals more protein peaks of low molecular sizes between 6 and 19 kDa, where not only a decrease, but also an increase of a protein peak with the molecular size of 10 kDa can be seen. Total nitrogen analysis revealed nitrogen levels of 1.734% for the unmalted, 1.725% for the germinating and 1.750% for the malted oats. According to Wu (1983), the total nitrogen content of oat kernels decreased at rst during germination, which is probably due to the loss of water-soluble nitrogen. After kilning the percentage of total nitrogen was increased. 3.3. Protein fractions Total protein analysis alone did not give a detailed picture of the changes occurring during the malting process. Proteins were fractioned according to Osborne to further investigate how oat protein alters during malting. 3.3.1. Albumin fraction The electropherogram of the water-soluble fraction of unmalted and malted oat samples using the Protein 80 LabChip is shown in Fig. 1A. Results indicate important changes among the overall malting process. After malting, new or increased protein peaks can be seen. The peak area reaching from 6 to 9 kDa and the peak at 61 kDa were not present before, but after malting. Increased protein peaks could be found at 13, 20 and 50 kDa. Proteins with molecular weights of 3544, 15 and 76 kDa were decreased during the process. Results from the Lab-on-a-Chip capillary electrophoresis agreed with the effect, found when two-dimensional gel electrophoresis (Fig. 1B) was applied. An increase in intensity of protein spots as well as new spots could be detected in the malted sample. Nevertheless, results of Lab-on-a-Chip analysis appear clearer and more precisely. 3.3.2. Globulin fraction Fig. 2A shows the patterns of the unmalted, germinating and malted oat globulin samples on the Protein 230 LabChip. Results reveal a reduction of the intensity of the group of bands located between 42 and 51 kDa (2) over the malting process. In the area of the molecular weights between 20 and 28 kDa (1), the band with the size of 25 kDa disappears. These patterns are clearly depicted in Fig. 2B, where the molecular weight distribution of the unmalted, germinating and malted globulin fraction can be seen in the electropherogram. A decrease in the protein peaks reaching from 42 to 51 kDa (2) was also detected. In this molecular weight range, protein peaks were degraded by 67% during the rst 3 days of germination and by 76% during the overall malting process, calculated based on the peak area. The electropherogram shows that the 25 kDa protein peak decreases in peak height by 30%, but
86
B
pH 3 90 kDa pH 10
10 kDa
Unmalted albumins Malted albumins

90 kDa
10 kDa
Fig. 1. (A) Electropherogram of the albumin fraction of unmalted and malted samples (Protein 80 LabChip). (B) 2D gels of unmalted (a) and malted (b) albumin fraction.
a decrease of the concentration of all the 2028 kDa proteins (1), calculated based on the peak area, could not be observed. Results of two-dimensional gel electrophoresis are shown in Fig. 2C. A large number of overlaying spots in the area of 4050 kDa in the unmalted sample (picture a) make the identication difcult. However, a decrease of these proteins after the malting process (picture c) can be seen. This fact concurs with the results from the Lab-on-a-Chip analysis, where a decrease in peak area 2 (42 51 kDa) could be detected. The lower molecular weight protein spots ranging from 20 to 30 kDa seem to be changed only slightly during the malting process, whereas the protein spots with pH of 56 decrease and the spot with pH 7 seems not to be most affected. This is in accordance with results of Lab-on-a-Chip analysis. 3.3.3. Prolamin (avenin) fraction The results from the Protein 230 LabChip of prolamin degradation during the malting process are illustrated in Fig. 3A. The electropherogram displays a low molecular weight peak area extending from 15 to 23 kDa (1), a protein peak with the size of 31 kDa (2) and an area between the molecular weights from 44 to 51 kDa (3). All prolamins appear to be decreased during the process. In area 1 a general decrease by about 90% could be observed. The peak height of the 31 kDa peak (2) was reduced during 3 days of germination from 188 to 107 FU and to 52 FU after kilning, what resembles a decline of 72%. The 4451 kDa part (3) was reduced from a peak area of 6837 FU during germination and to 13 FU after the entire malting process, which equals a decrease by 46 and 81%, respectively. Fig. 3B shows the gels of unmalted (a), germinating (b) and malted (c) prolamin samples. Protein spot areas of 15, 3035 and 4045 kDa could be observed. Results correspond approximately with the peak areas of the electropherogram. During germination the spot intensity decreases and, after kilning, almost all spots disappeared. 3.3.4. Glutelin fraction Glutelin proteins decreased during the malting process (Fig. 4A). A protein peak with molecular weight of 9 kDa could be detected,
which seems to increase slightly by 10% during malting. Peak areas ranging from 22 to 27 kDa and from 46 to 51 kDa were detected as well. A moderate decrease by 40% during the process could be measured at the protein peaks ranging from 20 to 27 kDa. The peaks between 46 and 51 kDa appear to be entirely degraded during the malting procedure. The acrylamide gels of the two-dimensional gel electrophoresis showed an overall decrease of intensity of spots after the malting process (Fig. 4B). 3.4. Amino acid analysis 3.4.1. Free and total amino acid composition of unmalted and malted oats In Table 1 the free amino acid composition of unmalted and malted oats is specied. Glutamic acid (254.1 mg/g oats) was found to be the major free amino acid in unmalted oats. In addition, aspartic acid (103.7 mg/g oats) and arginine (94.1 mg/g oats) were also present in high amounts in the unmalted oat kernel. Furthermore, no free proline could be detected. Results clearly show an increase in all free amino acids in oats after the malting process. Arginine (735.5 mg/g oats) was present in the highest concentration of malted oats, followed by proline (664.4 mg/g oats), threonine (654.3 mg/g oats), glutamic acid (481.6 mg/g oats) and phenylalanine (477.5 mg/g oats). Especially the free proline content increased from 0 to 664.4 mg/g oats during malting. And also tyrosine, threonine and phenylalanine increased signicantly by 98, 97, and 95%. Table 2 shows the total amino acid composition of unmalted and malted oats. Glutamic acid is by far the amino acid with the highest amount (20.61 mg/g) in unmalted oats. Asparitc acid, followed by arginine and leucine are also present in relatively high amounts of 8.27, 7.81 and 7.72 mg/g, respectively. Also the essential amino acids valine, phenylalanine and lysine could be found in high amounts of 5.86, 5.53 and 4.38 mg/g. Considering a maximum 5% error in the method used, the following amino acid levels signicantly decrease during malting: tyrosine (by 57%), arginine (by 16%), proline (by 14%), phenylalanine (by 14%) and glycine (by 13%). Only the amino acid histidin appears to increase during the process
87
A
90 kDa
C
pH 3 pH 10
a
40-50 kDa 20-30 kDa
10 kDa 90 kDa
2 1
B
1
10 kDa 90 kDa
2
Unmalted globulins Germinating globulins Malted globulins
10 kDa
Fig. 2. (A) SDS-PAGE image of the globulin fraction of unmalted, germinating and malted samples. (B) Electropherogram of the globulin fraction of unmalted, germinating and malted samples (Protein 230 LabChip). (C) 2D gels of unmalted (a), germinating (b) and malted (c) globulin fraction.
by 20%. However, in total, a value of 102.23 in the unmalted and a slightly lower level of 94.39 mg/g in the malted sample could be detected. 3.4.2. Total amino acid composition of unmalted, germinating and malted oat protein fractions The total amino acid composition of oat protein fractions shows signicant differences (Table 3). On the one hand, the amino acid composition changes during the malting process and on the other hand, amino acids occur in different amounts in the four fractions. In the albumin fraction, a total of 289.65 g amino acids/kg albumin could be found in the unmalted sample, whereas glutamic acid has the highest concentration (39.00 g/kg albumin), followed by glycine (24.96 g/kg albumin) and aspartic acid (24.09 g/kg albumin). All these amino acids were increased after malting by 12, 27 and 35%, respectively. These increases led to a total amino acid content in the malted sample of 357.60 g amino acids/kg albumin. For the unmalted globulin sample, a value of 587.18, for the germinating sample of 554.38 and, for the malted sample, 503.11 g amino acids/kg globulin was found. In the globulin fraction as well, glutamic acid has the highest, but in a 3.3-fold increased concentration (127.37 g/kg globulin). Aspartic acid (55.52 g/kg globulin),
arginine (54.27 g/kg globulin), leucine (44.03 g/kg globulin) and phenylalanine (34.30 g/kg globulin) appeared also in high concentrations. Except for lysine, which increased by 14%, all other amino acids decreased over the process. Glutamic acid was decreased by 28%, as well as tyrosine (by 22%), valine (by 20%) and serine (by 19%). For the prolamin fraction, a total of 354.07 g/kg prolamin was determined in the unmalted sample. The total concentration was decreased during germination to 311.69 g/kg prolamin and further decreased after kilning to 224.83 g/kg prolamin. Again, glutamic acid was found in the highest concentration (125.23 g/kg prolamin), but was decreased markedly during malting by 43%. Values of the other amino acids were a lot lower than glutamic acid concentration, but were degraded as well; leucine by 37%, valine by 38%, phenylalanine by 34% and tyrosine by 37%. The amino acid concentration of the glutelin fraction was (192.38 g/kg glutelin) in the unmalted sample and it was decreased during the process to 169.42 g/kg glutelin and further to 134.67 g/ kg glutelin. As already found in the other fractions, glutamic acid (37.76 g/kg glutelin) and aspartic acid (18.07 g/kg glutelin) were apparent in the highest concentrations, followed by arginine (16.46 g/kg glutelin), leucine (13.31 g/kg glutelin), valine (12.14 g/ kg glutelin) and phenylalanine (11.4 g/kg glutelin). All amino acids
88
B
90 kDa pH 3 pH 10
ca. 40-45 kDa ca. 30-35 kDa
A
10 kDa 90 kDa
ca. 15 kDa
Unmalted prolamins Germinating prolamins 3 Malted prolamins

10 kDa
90 kDa
10 kDa
Fig. 3. (A) Electropherogram of the prolamin fraction of unmalted, germinating and malted samples (Protein 230 LabChip). (B) 2D gels of unmalted (a), germinating (b) and malted (c) prolamin fraction.
were degraded markedly, especially proline (by 56%), tyrosine (by 41%) and valine (by 41%). 4. Discussion 4.1. Total protein analysis of oats The total nitrogen content of oat samples during malting was analysed using the combustion method. In this study a total nitrogen content of 1.734, 1.725 and 1.750% was found for the unmalted, germinating and malted sample. The protein content was calculated by multiplying the nitrogen content by the conversion factor 5.4 according to Mariotti et al. (2008). The calculated protein content is 9.36, 9.32 and 9.45% for the three samples. According to the literature (Lasztity, 1996) the nitrogen content of oats ranges from 11 to 15 %, probably determined by using conversion factors of 6.25 or 5.83. According to Mosse (1990) and Mariotti et al. (2008) this factor leads to overestimated protein content. Thus, the corrected conversion factor for oats of 5.4 was used for calculating the protein content of the oats used, which leads to the relatively low protein content of the analysed oat grain. A comparison of the free amino acid content in unmalted and malted oats (Table 1) revealed an increase in all amino acids during malting. This is due to hydrolysis of the native protein, which gives both high molecular weight and low molecular weight protein breakdown products (peptides and free amino acids) (Kunze, 1996). Quantitatively, glutamic acid and aspartic acid were most present in the unmalted sample. In the malted sample, instead arginine,
proline and threonine had the highest concentrations, which was also reported by Chittenden et al. (1987), where free proline was found to increase during germination of wheat. The largest increase could be observed in the concentration of free proline, which was not present in the unmalted sample and thus was increased by 100%. In particular, the essential amino acids isoleucine, leucine, lysine, threonine, valine and phenylalanine were highly increased by about 90% each. Total amino acid composition of unmalted oats (Table 2) corresponds to that reported by Lasztity (1996). There are no great differences during malting. Total amino acid concentration of unmalted oats is 102.23 and of malted grain is 94.39 mg/g. According to Briggs et al. (1981), there is no net loss or gain of nitrogen in barley grain during malting, apart from substances leached during steeping. However, tyrosine was decreased by 57% and histidine is the only amino acid, which was increased by 20% during malting. Before and after malting, glutamic acid and aspartic acid had the highest concentrations, which can be attributed to the fact that during hydrolysis glutamine and asparagine were converted into their acids. 4.2. Protein fractions 4.2.1. Albumin fraction The albumin fraction of other cereals, such as barley, contains metabolically active proteins (Lasztity, 1996); thus, most of the enzymes of oats are probably found in the water-soluble fraction as well. Since the main purpose of malting is the production of
89
B A
pH 3 90 kDa pH 10
Unmalted glutelins 2 1 Malted glutelins 3
10 kDa 90 kDa
10 kDa
Fig. 4. (A) Electropherogram of the glutelin fraction of unmalted and malted samples (Protein 80 LabChip). (B) 2D gels of unmalted (a) and malted (b) glutelin fraction.
enzymes, an increase in the albumin fraction was expected. As shown in Fig. 1 and Table 3, an increase in several proteins could be observed. Proteins with molecular weights of 1417, 2027 and 36 47 kDa have been described in the literature (Lasztity, 1996) and could be detected with the Protein 80 LabChip, whereas the 36 47 kDa protein is decreasing during malting. The other proteins (13, 20 and 2326 kDa) were increased during the process, which can be explained by the increasing enzyme activity. In addition, a smaller protein peak area (69 kDa) was found with LabChip, which cannot be detected with gel electrophoresis, because the commonly used protein analysis technique detects proteins above 14 kDa only. These small proteins and the protein with the size of 61 kDa are probably protein breakdown products, which are formed during malting and are water soluble. The formation of water-soluble protein breakdown products is well known for barley proteins (Kunze, 1996). The isoelectric points of oat albumins, detected with two-dimensional gel electrophoresis (Fig. 1B), range from pH 5 to 8. Similar values have been reported by Ma and Harwalkar (1984). Amino acid analysis of oat albumins revealed an
increase in all amino acids, with the exception of arginine and histidine. The increase of proline by 48% is notable. The high increase in albuminproline probably contributes to the strong increase in free proline as mentioned in Section 4.1. But also phenylalanine, aspartic acid and tyrosine were increased markedly. This effect concurs with results reported by Wu (1983), where an increase in phenylalanine plus tyrosine was reported during germination. 4.2.2. Globulin fraction The salt water-soluble protein of oats is known to be the major protein fraction. This fraction functions as storage proteins and consists of 3S, 7S and 12S globulins (Lasztity, 1996; Shewry and Halford, 2002). Peterson (1978) showed that the 12S globulin has a molecular weight of 322 kDa and consists of two subunits with molecular weights of about 32 and 22 kDa, called the a- and bsubunits. As shown in Fig. 2, these subunits could be detected with the Protein 230 LabChip, covered in area 1. This type of protein seems to be affected only slightly by the degradation process
Table 1 Free amino acid composition of the unmalted and malted oat grain (mg/g [dry wt]) Amino Acid Essential Isoleucine Leucine Lysine Threonine Valine Phenylalanine Histidin Arginine Nonessential Alanine Aspartic acid Glutamic acid Glycine Proline Serine Tyrosine Unmalted 15.0 0.7 18.5 0.9 42.3 2.0 21.2 1.0 46.2 2.1 22.1 1.1 58.0 2.8 94.1 4.0 55.0 2.2 103.7 5.0 254.1 5.2 21.4 1.0 0 25.4 1.2 8.0 0.3 Malted 245.3 5.0 316.5 9.0 356.5 9.5 654.3 12.0 430.3 10.1 477.5 9.9 322.5 8.9 735.5 13.2 364.5 9.3 189.1 4.5 481.6 10.5 67.2 2.7 664.4 12.1 277.3 6.4 336.9 6.9
Table 2 Total amino acid composition of the unmalted and malted oat grain (mg/g [dry wt]) Amino Acid Essential Isoleucine Leucine Lysine Threonine Valine Phenylalanine Histidin Arginine Nonessential Alanine Aspartic acid Glutamic acid Glycine Proline Serine Tyrosine Unmalted 3.88 0.2 7.72 0.4 4.38 0.2 3.22 0.1 5.86 0.2 5.53 0.2 3.00 0.1 7.81 0.4 5.31 0.2 8.27 0.4 20.61 1.0 5.56 0.2 5.94 0.3 4.43 0.2 2.87 0.1 Malted 3.79 0.1 7.16 0.3 4.36 0.2 3.34 0.1 5.17 0.2 4.77 0.2 3.59 0.1 6.57 0.3 4.79 0.2 8.89 0.4 19.38 0.9 4.85 0.2 5.09 0.2 4.44 0.2 1.25 0.0
90
Table 3 Average amino acid compositions of the different protein fractions extracted from oats grown in Finland (mg/g protein fraction) Albumin 1a Isoleucine Leucine Lysine Threonine Valine Phenylalanine Histidin Arginine Alanine Aspartic acid Glutamic acid Glycine Proline Serine Tyrosine Total 8.33 0.4 17.45 0.8 13.91 0.6 11.54 0.5 12.13 0.6 9.73 0.4 8.15 0.4 16.88 0.8 17.52 0.8 24.09 1.2 39.00 1.9 24.96 1.2 13.45 0.6 20.75 1.0 12.44 0.6 289.65 Prolamin 1c Isoleucine Leucine Lysine Threonine Valine Phenylalanine Histidin Arginine Alanine Aspartic acid Glutamic acid Glycine Proline Serine Tyrosine Total 11.44 0.5 33.34 1.6 4.84 0.2 7.51 0.3 18.66 0.9 21.73 1.0 9.51 0.4 17.13 0.8 14.19 0.7 12.72 0.6 125.23 4.2 8.26 0.4 22.51 1.1 9.18 0.4 4.60 0.2 354.07 2c 10.48 0.5 29.87 1.4 5.01 0.2 7.02 0.3 16.69 0.8 18.53 0.9 8.77 0.4 16.36 0.8 13.04 0.6 12.17 0.6 104.82 5.0 7.93 3.9 19.97 0.9 8.31 0.4 4.06 0.2 311.69 3c 7.48 0.3 21.15 1.0 4.03 0.2 5.21 0.2 11.54 0.5 14.33 0.7 7.54 0.3 11.23 0.5 9.81 0.4 10.91 0.5 71.79 3.5 6.24 0.3 15.35 0.7 6.30 0.3 2.89 0.1 224.83 2a 7.80 0.3 16.45 0.8 11.88 0.5 11.11 0.5 11.18 0.5 11.68 0.5 6.92 0.3 13.33 0.6 15.41 0.7 24.07 1.2 35.07 1.7 23.90 1.1 14.77 0.7 20.60 1.0 12.65 0.6 276.88 3a 10.03 0.5 20.74 1.0 15.41 0.7 14.59 0.7 14.61 0.7 13.18 0.6 8.00 0.4 15.73 0.7 20.82 1.0 32.46 1.6 43.84 2.1 31.57 1.5 19.92 0.9 26.74 1.3 16.64 0.8 357.60 Globulin 1b 26.49 1.3 44.03 2.2 18.89 0.9 20.85 1.0 33.34 1.6 34.30 1.7 18.16 0.9 54.27 2.7 27.64 1.3 55.52 2.7 127.37 6.3 28.81 1.4 23.09 1.1 29.97 1.4 18.84 0.9 587.18 Glutelin 1d 8.96 0.4 13.31 0.6 7.50 0.3 6.55 0.3 12.14 0.6 11.40 0.5 7.93 0.3 16.46 0.8 9.05 0.4 18.07 0.9 37.76 1.8 8.82 0.4 6.76 0.3 8.89 0.4 6.58 0.3 192.38 2d 8.63 0.4 13.43 0.6 3.74 0.1 5.85 0.2 11.00 0.5 10.53 0.5 6.67 0.3 15.00 0.7 8.55 0.4 17.14 0.8 32.00 1.5 8.03 0.4 5.81 0.2 7.48 0.3 5.79 0.2 169.42 3d 6.21 0.3 10.37 0.5 4.79 0.2 4.65 0.2 7.21 0.3 9.81 0.4 7.37 0.3 12.08 0.6 7.36 0.3 12.31 0.6 27.48 1.3 6.17 0.3 2.98 0.1 5.97 0.2 3.85 0.1 134.67 2b 26.25 1.3 42.09 2.1 19.40 0.9 19.66 0.9 29.71 1.4 31.50 1.5 17.92 0.8 52.31 2.6 27.08 1.3 52.93 2.6 114.37 5.7 28.63 1.4 24.46 1.2 26.93 1.3 16.85 0.8 554.38 3b 23.21 1.1 38.97 1.9 21.53 1.0 19.04 0.9 26.68 1.3 29.25 1.4 17.34 0.8 44.96 2.2 27.53 1.3 50.55 2.5 91.51 4.5 26.83 1.3 21.39 1.0 24.32 1.2 14.72 0.7 503.11
Sample 1a1d unmalted; sample 2a2d germinating; sample 3a3d malted.
taking place during malting. In contrast to the 12S globulin, the 7S globulin, which mainly consists of a 55 kDa polypeptide and is found in area 2 in Fig. 2, is decreasing during the malting process. This area probably represents only the 7S globulin, although the disulde-bonded dimer, formed by the a- and b-subunit has a molecular weight of 54 kDa. Due to the fact that analysis was performed under reducing conditions, no disulde-bond remained. The 3S fraction was described in the literature as a protein consisting of a 15 and 21 kDa subunit (Lasztity, 1996). The 15 kDa polypeptide could not clearly be detected with the LabChip, but the 21 kDa component was probably found in area 1 (Fig 2B). Amino acid analysis revealed decreasing proteins as well, since the total amino acid amount of the globulin fraction is decreasing, which concurs with the degradation in the globulin fraction during malting. In particular the lysine content was found to be increased during the process, which is remarkable from a nutritional point of view and was also found by Wu (1983), who found an increasing lysine content in germinating oats. This can be attributed to the fact that the globulin fraction, which is high in lysine, was degraded during the malting process. This probably revealed free lysine, which is water soluble and thus, was extracted with the albumin and globulin fraction. The amino acid composition of globulins corresponds to that reported from Lasztity (1996). Glutamic acid is the most frequently found amino acid and a decrease by 28% of this amino acid was detected, which is probably due to decreases in the 12S globulin subunit, since the 12S subunit has been described to contain a higher amount of glutamic acid than the 7S globulin.
4.2.3. Prolamin fraction Since oats main storage protein fraction is the globulin fraction, oats prolamin fraction only accounts for 15% of total oat protein (Lasztity, 1996), which is different compared to other cereals, such as wheat, barley or rye. However, two-dimensional gel electrophoresis revealed more than 20 components (Fig. 3B, a). The Labon-a-Chip analysis disclosed three main peak areas in the same molecular weight range (Fig. 3A). The pH of isoelectric focusing was found to range from pH 6 to 9. These results concur with those reported by Lasztity (1996). During the malting process the prolamin fraction was degraded almost entirely, which is in agreement with other germinating cereals such as barley, reported by Celus et al. (2006). The degradation by endoproteases is due to the fact that storage proteins are supporting the embryo during the rst stages of development. The oats prolamin fraction is like many other cereals also rich in the amino acids proline and glutamine (Simpson, 2001). Thus, prolamin proteins most likely function as storage proteins, because the aim of the glutamine- and prolinerich polypeptides is to supply the embryo with amino acids and nitrogen during germination. 4.2.4. Glutelin fraction The proteins remaining after removing the water-soluble albumins, the salt water-soluble globulins and the alcohol soluble prolamins are called glutelins. Unfortunately, these extractions are generally incomplete and some nitrogen remains in the residue. The quantity of glutelins observed is often directly related to the efciency of the preceding extractions of the albumins, globulins
91
and prolamins (Shewry et al., 1978). Thus, it is likely to extract not only glutelins with the last extraction step, but also remaining proteins from the water, salt water and alcohol soluble fractions, especially the globulin fraction (Lasztity, 1996). Robert et al. (1985) reported two-dimensional gel electrophoresis of oats proteins, where they found a few proteins spots, which probably are true candidates for the glutelin fraction, along with proteins spots, which are most likely a- and b-globulin subunits. A similar effect can be seen in our results of the glutelin fraction (Fig. 4). Compared to the gels from the globulin fraction (Fig. 2C) it seems likely that not all globulins had been extracted with the 5.0% NaCl solution and remained until the extraction with urea, SDS and DTT, since the molecular weights of 2227 and 4251 kDa could be found in both fractions. The proteins ranging from 22 to 27 kDa are probably bsubunit proteins from the 12S globulin. However, the LabChip revealed a so far unknown polypeptide with molecular weight of 9 kDa, which is most likely a glutelin protein, that seems to be unaffected by the malting process. In conclusion, this study reveals an understanding about the protein changes taking place during malting of oats. In general a degradation of the proteins to small peptides and amino acids could be observed in any fraction except in the albumin fraction, in which some proteins increased in amount. This is due to the fact that these represent most likely metabolically active proteins. Amino acid analysis supported the observation of increased protein amount in the albumin fraction and decreased protein amounts in the other fractions. In addition, so far unknown proteins could be detected in the albumin and glutelin fraction with the Lab-on-a-Chip analysis. This technique was found to be appropriate for analysis of degrading or increasing proteins, as it revealed repeatable and reliable results, which could be validated by using common protein analysis techniques such as two-dimensional gel electrophoresis. Acknowledgements This project was supported by the Irish Government under the National Development Plan, 20062013. The authors would like to thank Mrs. Paula OConnor for her support in amino acid analysis. References
Agilent Technologies, 2005. Agilent 2100 Bioanalyzer - 2100 Expert Users Guide. Agilent Technologies, Waldbronn, Germany. Briggs, D.E., Hough, J.S., Stevens, R., Young, T.W., 1981. Malting and Brewing Science, second ed. Chapman & Hall, London, U.K. Brijs, K., Trogh, I., Jones, B.L., Delcour, J.A., 2002. Proteolytic enzymes in germinating rye grains. Cereal Chemistry 79, 423428. Celus, I., Brijs, K., Delcour, J.A., 2006. The effects of malting and mashing on barley protein extractability. Journal of Cereal Science 44, 203211.
Chittenden, C.G., Laidman, D.L., Ahmad, N., Wyn Jones, R.G., 1987. Amino acids and quaternary nitrogen compounds in the germinating wheat grain. Phytochemistry 17, 12091216. Kunze, W., 1996. Technology Brewing and Malting, seventh ed. Versuchs- und Lehranstalt fur Brauerei, Berlin, Germany. Lapvetelainen, A., Bietz, J.A., Huebner, F.R., 1995. Reversed-phase high-performance liquid chromatography of oat proteins: application to cultivar comparison and analysis of the effect of wet processing. Cereal Chemistry 72, 259264. Lasztity, R., 1996. The Chemistry of Cereal Proteins, second ed. CRC Press, Boca Raton, Florida. Ma, C.-Y., Harwalkar, V.R., 1984. Chemical characterisation and functionality assessment of oat protein fractions. Journal of Agricultural and Food Chemistry 32, 144149. Mariotti, F., Tome, D., Mirand, P.P., 2008. Converting nitrogen into protein beyond 6.25 and Jones factors. Critical Reviews in Food Science and Nutrition 48, 177184. Moore, S., Stein, W.H., 1963. Chromatographic determination of amino acids by the use of automatic recording equipment. Methods in Enzymology 6, 819831. Mosse, J., 1990. Nitrogen to protein conversion factor for ten cerals and six legumes or oilseeds. A reappraisal of its denition and determination. Variation according to species and to seed protein. Journal of Agricultural and Food Chemistry 38, 1824. Peraaho, M., Kaukinen, K., Mustalahti, K., Vuolteenaho, N., Maki, M., Laippala, P., Collin, P., 2004. Effect of an oats-containing gluten-free diet on symptoms and quality of life in coeliac disease. A randomized study. Scandinavian Journal of Gastroenterology 39, 2731. Peterson, D.M., 1978. Subunit structure and composition of oat seed globulin. Plant Physiology 62, 506509. Peterson, D.M., 2001. Oat antioxidants mini review. Journal of Cereal Science 33, 115129. Ren, Y., Ellis, P.R., Ross-Murphy, S.B., Wang, Q., Wood, P.J., 2003. Dilute and semidilute solution properties of (1 / 3), (1 / 4)-b-D-glucan, the endosperm cell wall polysaccharide of oats (Avena sativa L.). Carbohydrate Polymers 53, 401408. Robbins, G.S., Pomeranz, Y., Briggle, L.W., 1971. Amino acid composition of oat groats. Journal of Agricultural and Food Chemistry 19, 536. Robert, L.S., Nozzolillo, C., Altosaar, I., 1985. Characterization of oat (Avena sativa L.) residual proteins. Cereal Chemistry 62, 276279. Robert, L.S., Matlashewski, G.J., Adeli, K., Nozzolillo, C., Altosaar, I., 1983a. Electrophoretic and developmental characterization of oat (Avena sativa L.) globulins in cultivars of different protein content. Cereal Chemistry 60, 231234. Robert, L.S., Nozzolillo, C., Altosaar, I., 1983b. Molecular weight and charge heterogeneity of prolamins (avenins) from nine oat (Avena sativa L.) cultivars of different protein content and from developing seeds. Cereal Chemistry 60, 438442. Shewry, P.R., 1995. Plant storage proteins. Biological Reviews 70, 375426. Shewry, P.R., Halford, N.G., 2002. Cereal seed storage proteins: structures, properties and role in grain utilization. Journal of Experimental Botany 53, 947958. Shewry, P.R., Hill, J.M., Pratt, H.M., Leggatt, M.M., Miin, B.J., 1978. An evaluation of techniques for the extraction of hordein and glutelin from barley seed and a comparison of the protein composition of Bomi and Ris 1580. Journal of Experimental Botany 29, 677692. Simpson, D.J., 2001. Proteolytic degradation of cereal prolamins the problem with proline. Plant Science 161, 825838. Wu, Y.V., 1983. Effect of germination on oats and oat protein. Cereal Chemistry 60, 418420. Zarkadas, C.G., 1982. A comparison of the amino acid composition of two commercial oats groats. Cereal Chemistry 59, 323327.

Effect of milling, pasta making and cooking on minerals in durum wheat

Francesco Cubadda a, *, Federica Aureli a, Andrea Raggi a, Marina Carcea b
a b
` Istituto Superiore di Sanita, Viale Regina Elena 299, 00161 Rome, Italy Istituto Nazionale di Ricerca per gli Alimenti e la Nutrizione, Via Ardeatina 546, 00178 Rome, Italy
Article history: Received 14 April 2008 Received in revised form 8 July 2008 Accepted 9 July 2008 Keywords: Durum wheat Pasta Minerals Trace elements Milling Processing Cooking
a b s t r a c t
The effect of technological processing on the contents of eight minerals i.e., calcium, copper, iron, magnesium, phosphorous, potassium, selenium, and zinc was investigated in pasta making. Milling of durum wheat as well as pasta making were carried out in a pilot plant by using three different grain samples. Pasta samples purchased on the market were also surveyed to gain information on the mineral content of commercial products. The effect of cooking was also investigated in order to determine the retention of the selected elements in the nal ready-to-eat product. Analyte concentrations in whole grains, semolina, pasta and cooked pasta were determined by inductively coupled plasma-mass spectrometry. Conventional roller milling signicantly reduced the content of each mineral in durum wheat grains. However concentration losses as a consequence of milling widely differed among elements, from 16% for Se to 66% for Mg and Zn on a dry weight basis. Retention of elements after milling followed the order Se > Ca > Cu > P z K > Fe > Mg z Zn. Pasta making had little effect on element concentrations in semolina. Cooking caused an increase in the calcium content of pasta whereas the concentrations of the other elements were either unchanged or slightly reduced (018% on a dry weight basis) except potassium, which showed a decrease of 74%. Commercial pasta samples showed concentrations of minerals similar to those of the experimental samples, except selenium which was higher due to the use of imported wheat with higher levels of selenium in industrial semolina production. Overall, pasta appears to be a valuable source of several minerals, especially selenium, copper, magnesium, and zinc. 2008 Elsevier Ltd. All rights reserved.
1. Introduction Cereals and derived products are among the major dietary sources of essential elements for humans. The contribution of cereal products to the estimated dietary intake of several minerals and nutritionally benecial trace elements is about 2030% of the total intake in Western countries (Carcea et al., 2007). In the case of iron and manganese the contribution is as high as 4050% (Carcea et al., 2007). These gures are even higher in other regions of the world (Choi and Kim, 2007; Hattori et al., 2004) and especially in developing countries (Hussein and Bruggemann, 1997). Wheat is the main cereal crop used for human consumption in many areas worldwide. Common wheat (Triticum aestivum L.) is widely used for breadmaking, whereas durum wheat, i.e., Triticum turgidum L. subsp. durum (Desf.) Husn., is mainly employed in the production of other food items, pasta being the most popular. In Italy, the main pasta producer, pasta is a staple food. However pasta
Abbreviations: RDA, daily recommended dietary allowance. * Corresponding author. Tel.: 39 06 4990 3643; fax: 39 06 4990 2540. E-mail address: francesco.cubadda@iss.it (F. Cubadda). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.07.008
consumption has become widespread in several countries (UNIPI, 2007). Notwithstanding its popularity, the mineral prole of pasta as inuenced by processing (i.e., milling and pasta making) and cooking has been the subject of systematic investigations mainly in the 1980s (Albrecht et al., 1987; Brondi et al., 1984; Ranhotra et al., 1984, 1985; Toepfer et al., 1972; Yaseen, 1993). The introduction of new cultivars as well as alterations in agronomic practices and changes in the processing conditions do require updated studies. These studies should employ state-of-the-art techniques and include analytical quality control schemes allowing, for instance, precise and accurate determinations of essential microelements such as selenium. Milling is the critical process affecting the concentrations of inorganic elements in wheat-derived food products. As the outer parts of the kernel, especially the aleurone layer and the germ, are richer in minerals when compared to the starchy endosperm, conventional milling reduces their content in our (semolina, in the case of durum wheat) and concentrates them in the milling residues (Brondi et al., 1984). However, it has been reported that differences in the mineral content likely exist even between the inner endosperm and the outer endosperm (Pomeranz, 1988). The grain shape and texture (which both depend also on cultivar) and
F. Cubadda et al. / Journal of Cereal Science 49 (2009) 9297
93
the technical conditions of milling, primarily the extraction rate, are important in determining the extent of mineral loss. However, when all these variables are xed, the distribution of the mineral in the various milling fractions ultimately depends on how the element is unevenly distributed within the kernel. Therefore, it will vary on an element-specic basis. There is little information on the effect of milling on durum wheat minerals (Brondi et al., 1984; Toepfer et al., 1972). This information is needed to assess the nutritional signicance of mineral loss as a consequence of conventional roller milling and to provide basic knowledge in order to establish, for instance, whether there is a need for fortication, which minerals should be supplemented (in bioaccessible form), and at what levels. Limited and sometimes contradictory data are available on the effect of further processing of semolina into pasta as well as pasta cooking (Albrecht et al., 1987; Ranhotra et al., 1984, 1985; Yaseen, 1993). The present study was undertaken to determine the effect of conventional roller milling, pasta making and cooking on eight nutritionally important minerals (namely calcium, copper, iron, magnesium, phosphorous, potassium, selenium, and zinc) in durum wheat. Milling and pasta making pilot plants were employed for the manufacture of semolina and long (spaghetti type) pasta in order to be able to control the whole processing chain. Commercial pasta samples from major Italian brands were purchased on the market and included in this study for comparison. The effect of cooking was investigated in order to gain information on the retention of the selected elements in the nal ready-to-eat product. 2. Experimental 2.1. Milling and pasta making Three samples of durum wheat grain of about 10 kg each were included in the study. The samples belonged to three different cultivars, namely, Appio, Duilio, and Simeto (hereafter indicated as samples 1, 2, and 3) grown in Southern Italy and widely used in Italy for industrial pasta production. Semolina was obtained from each sample by cleaning and tempering the grains for 3640 h to 16.5% moisture and then milling them in a pilot mill (Model MLU 202, Buhler, Uzwill, Switzerland) equipped with three breaks, three reduction rolls, six steel screens and coupled with a purier to give standard grade semolina, following Approved Methods 26-10 A and 26-41 (AACC, 2000). Semolina yield (ash <0.90% d.m.) was about 66% for the three samples, and granulation was within Italian legal limits (25% maximum passed through a 180 mm sieve). Semolina was used to manufacture long pasta (spaghetti) by means of a pilot pasta making apparatus (Pavan, Padova, Italy) composed of a press and a dryer. The press (capacity 520 kg) was equipped with a vacuum mixing and extruding system, and a watercooling jacket on the barrel and the extrusion head to reduce heat and to maintain a constant extrusion temperature of no more than 50 C. Semolina was mixed with tap water for about 15 min to obtain a dough suitable for extrusion. The moisture content of the dough was about 30%. After extrusion, spaghetti were dried according to a 27 h drying cycle during which relative humidity was kept under control and maximum drying temperature was 58 C for a period of 4 h. The diameter of dry spaghetti was 1.651.70 mm. At the end of the drying cycle pasta was conditioned at room temperature (20 C) for 24 h and packed in polyethylene bags. 2.2. Cooking
(0.68 meq L1 hardness) with low levels of the elements being investigated, was used for all cooking. Common kitchen salt taken from a single commercial packet was added at a concentration of 9 g L1. The analyte levels were determined in water samples before cooking and were constant throughout the study. The optimum cooking time was taken as that required for the white core in the strands to disappear when they were squeezed between two test glasses (AACC, 2000). Cooked samples were drained in a plastic drainer to avoid contamination and aliquots of pasta strands were placed in suitable containers for moisture determination. 2.3. Sample preparation for analytical determinations Sample handling was carried out under clean room conditions. Tap water used in dough preparation and for cooking was acidied soon after sampling with ultrapure HNO3 (1% v/v) and stored at 4 C pending analysis. Grain and pasta samples were ground in order to obtain a ne our. As a previous study demonstrated that the use of common laboratory mills may lead to sample contamination at this step (Cubadda et al., 2001), an automatic pestle mill with internal parts made in agate model RM100 by Retsch GmbH & Co. (Haan, Germany) was used. A preliminary study on wheat grain samples milled by means of this device indicated the absence of any detectable contamination for the elements investigated in the present study. Aliquots of about 100 g of grain and dry pasta samples were ground. For each sample, grinding was carried out twice and the our obtained in the rst run was discarded. After each batch, the mortar was thoroughly cleaned with laboratory wipers and then rinsed with ultrapure water. This procedure was adopted in order to avoid carry-over contamination from preceding samples. In addition to experimental samples, 12 commercial pasta samples from eight different makers were identically ground and subsequently submitted to the same analytical procedure. Cooked and drained aliquots of pasta were similarly homogenized before being submitted to elemental analysis. Transfer of samples from one container to another was always performed with suitable plastic tools in order to avoid contamination. Sample digestion was performed by means of a closed vessel system, a Milestone MLS-1200 Mega MW oven (FKV, Bergamo, Italy), as described elsewhere (Cubadda et al., 2002). The MW system utilized was able to handle pressures as high as 110 bar. Sample weight was 0.30.4 g on a dry weight basis, depending on the material. The reagents used in sample digestion were ultrapure concentrated HNO3 (Carlo Erba Reagenti, Milan, Italy) and H2O2 (Merck, Darmstadt, Germany), 3 and 1 ml respectively. Each sample was digested in duplicate and made up to 30 ml in polypropylene disposable tubes with high purity deionized water (18.2 MU cm resistivity at 25 C) obtained by a Milli-Q apparatus (Millipore, Molsheim, France). Dry weight determinations were performed for each sample on separate 10-g aliquots by oven drying at 103 C until constant weight. Two reference materials were included randomly in analytical batches from digestion onward. These were the RM 8436 (durum wheat our), from the National Institute of Standards and Technology (Gaithersburg, MD), and the BCR CRM 189 (wholemeal our), from the Institute for Reference Materials and Measurements (Geel, Belgium). 2.4. Instrumentation and analytical measurements
Experimental pasta samples were cooked in Pyrex asks previously decontaminated with a 12 h treatment with nitric acid (ultrapure, 10% v/v solution). A single batch of soft tap water
A quadrupole Sciex Elan 6000 ICP mass spectrometer (PerkinElmer, Norwalk, CT, USA), equipped with an ASX-500 autosampler
94
(model 510) and an ADX-500 autodilutor (both from CETAC Technologies, Omaha, NE, USA), was used for quantication of the analytes. Details of the instrumentation and the operating conditions are reported elsewhere (Cubadda et al., 2002). The analytical masses were 43Ca, 63Cu, 57Fe, 39K, 24Mg, 31P, 7782Se, and 64Zn. Multielemental calibration standards were prepared from 1000 mg L1 stock solutions of individual elements (BDH, Poole, England) by dilution with 3% v/v ultrapure concentrated HNO3. In all measurements, rhodium (20 mg L1) was selected as internal standard for correction of matrix effects and instrumental drift. Correction of 44Ca16OH interference on 57Fe was accomplished by application of a mathematical equation calculated daily as reported previously (Cubadda et al., 2002). Randomly selected samples were analysed on different days to verify the precision of results. Samples of the two reference materials were analysed in each analytical run to check accuracy of measurements. Digestion blanks were run together with samples belonging to the same analytical batch and their signal was subtracted from that of the sample when calculating analyte concentrations. 2.5. Statistical analysis The existence of signicant differences in the element concentrations (on a dry weight basis) among the studied matrixes on account of the various treatments (i.e., milling, pasta making, cooking) was determined by analysis of variance. The test resulted signicant for all metals (p 0.05) and thus a multiple comparison test (Scheffe test) was performed to identify which treatments caused signicant variations (p 0.05). 3. Results and discussion The results obtained for the reference materials analysed for quality control purposes are summarized in Table 1. Good agreement was observed with the certied or best estimated values of each element, indicating effective recovery of analytes after digestion and subsequent accurate detection. Precision expressed as the coefcient of variation (CV) was, on average, 6% for calcium, 3% for copper, 2% for iron, 1% for potassium, 1% for magnesium, 2% for phosphorous, 2% for selenium, and 4% for zinc. Tables 24 show the element concentrations measured in the three grain samples selected for this study and in their derived products, including cooked pasta. Each result is the average of four experimental replicates, on which duplicate subsampling was carried out for analyses. Data are reported both on a dry and on a fresh weight basis. Average moisture content of samples was 9.8%,
12.6%, 10.1% and 61.1% for grain, semolina, pasta and cooked pasta, respectively. In order to allow easy recognition of the effects of processing and cooking, the overall percent variation in element concentrations caused by the different treatments in relation to their original levels in grain (100%) is shown in Tables 24, together with the percent variation of the element concentrations in each type of sample as a consequence of a specic treatment. 3.1. Effect of milling As expected, milling signicantly reduced the concentrations of all elements (p < 0.01), with average losses ranging from 16% for selenium to 66% for magnesium and zinc on a dry weight basis (Tables 24). A modest reduction of selenium concentration upon milling (8%20%) has also been found in studies on common wheat (Lyons et al., 2005; Toepfer et al., 1972). Similar to sulphur, selenium, which occurs predominantly as selenomethionine in wheat grains, is mostly protein-bound and more evenly distributed throughout the kernel when compared to other minerals (Lyons et al., 2005). Therefore, due to a higher proportion being stored in the endosperm, lower proportions of Se are removed in the milling process. Calcium showed an average loss of 41% in concentration terms, which is in agreement with the results obtained in earlier work on durum wheat (Toepfer et al., 1972). Studies on the distribution of minerals within the kernel of common wheat reported that the endosperm contains about 50% of calcium whereas about 25% is found in the aleurone layer (Pomeranz, 1988). For other elements, such as magnesium and zinc, only a minor proportion is found in the endosperm, whereas !70% and w50%, respectively, is in the aleurone (Bock, 2000; Pieczonka and Rosopulo, 1985; Pomeranz, 1988). Accordingly, calcium was found to follow an entirely different pattern from that of magnesium, zinc and iron in industrially milled wheat ours (Lorenz et al., 1980). The level of calcium depended mainly on the calcium content of the wheat, whereas magnesium, zinc and iron depended more on milling variables (Lorenz et al., 1980). Next to calcium, copper showed a concentration decrease of 47% in this study. This is similar to the results obtained in earlier work on durum wheat (Brondi et al., 1984; Toepfer et al., 1972) and matches studies showing that about 45% of copper is in the endosperm of common wheat kernels (ODell et al., 1972; Pieczonka and Rosopulo, 1985). The concentration of phosphorous and potassium in semolina was 56% lower than that in parent grains. Phosphorous mainly occurs in wheat kernels as phytic acid and its salts. Potassium is associated with phosphorous as it forms, together with magnesium, a major part of the phytates found in the kernels (Bock, 2000). A sizeable reduction of phytic acid in semolina is achieved upon milling of durum wheat grains so that a major proportion of phosphorous is present as nucleoprotein, lipid, and inorganic phosphorous (Pomeranz, 1988). Since phytic acid is a metalchelating agent which can lower the absorption of several essential metals (including iron, zinc, calcium and magnesium) its removal into milling by-products is nutritionally benecial (Bock, 2000; ODell et al., 1972). Recently, low phytic acid mutants of wheat have been isolated with the view of improving the nutritional quality of wheat by reducing the major storage form of phosphorous and increasing the level of inorganic phosphorous, which is more readily absorbed by humans and other monogastric animals (Guttieri et al., 2006). A group of three elements showed a major decrease in concentration following milling, i.e., iron (63%), magnesium and zinc (66%). These metals are particularly abundant in the aleurone (ODell et al., 1972; Pieczonka and Rosopulo, 1985). Zinc is present
Table 1 Results obtained for the reference materials analysed (N 6) Element CRM 189 (wholemeal our) Found Mean Ca Cub Feb Kd Mgd Pd Seb Znb
a b c d b
RM 8436 (durum wheat our) Found Mean 276 4.28 41.1 3.20 1.12 2.91 1.25 22.3 (c.i.)a (14) (0.10) (1.2) (0.02) (0.02) (0.06) (0.02) (1.1) Best estimated Mean 278 4.30 41.5 3.18 1.07 2.90 1.23 22.2 (c.i.)a (26) (0.69) (4.0) (0.14) (0.08) (0.22) (0.09) (1.7)
Certied (c.i.)a (22) (0.13) (0.6) (0.09) (0.01) (0.03) (0.002) (1.1) Mean [520] 6.4 68.3 [6.3]c [1.9]c [5.3]c 0.132 56.5
c
(c.i.)a (0.2) (1.9)
540 6.36 69.2 6.23 1.92 5.49 0.137 56.6
(0.010) (1.7)
Uncertainty as half-width of the 95% condence interval of the mean. Concentrations in mg g1. Indicative value. Concentrations in mg g1.
F. Cubadda et al. / Journal of Cereal Science 49 (2009) 9297 Table 2 Variations of potassium, phosphorous, and magnesium concentration in durum wheat as a consequence of processing and cooking Sample On a fresh weight basis Grain Potassium 1a 2a 3a % vs precb % vs grainc Phosphorous 1a 2a 3a % vs precb % vs grainc Magnesium 1a 2a 3a % vs precb % vs grainc
a b c
95
On a dry weight basis Pasta 2.06 0.02 1.87 0.01 2.14 0.02 104 (103104) 44 (4247) 1.76 0.03 1.73 0.03 1.74 0.03 103 (103103) 44 (4347) 0.40 0.01 0.38 0.01 0.41 0.01 104 (104105) 35 (3437) Cooked pasta 0.22 0.01 0.20 0.01 0.25 0.01 11 (1112) 5 (55) 0.65 0.01 0.61 0.01 0.59 0.01 36 (3437) 16 (1517) 0.18 0.01 0.16 0.01 0.18 0.01 43 (4344) 15 (1516) Grain 4.85 0.04 4.79 0.04 5.62 0.05 100 100 4.48 0.07 4.10 0.07 4.50 0.07 100 100 1.33 0.01 1.15 0.01 1.33 0.01 100 100 Semolina 2.28 0.02 2.06 0.02 2.37 0.02 44 (4247) 44 (4247) 1.95 0.03 1.92 0.03 1.93 0.03 44 (4347) 44 (4347) 0.44 0.01 0.42 0.01 0.45 0.01 34 (3336) 34 (3336) Pasta 2.30 0.02 2.08 0.02 2.37 0.02 101 (100101) 44 (4247) 1.96 0.03 1.92 0.03 1.94 0.03 100 (100100) 45 (4347) 0.45 0.01 0.43 0.01 0.46 0.01 101 (101102) 35 (3437) Cooked pasta 0.58 0.01 0.52 0.01 0.64 0.01 26 (2527) 11 (1112) 1.68 0.03 1.57 0.03 1.52 0.02 82 (7986) 37 (3438) 0.45 0.01 0.42 0.01 0.46 0.01 100 (99101) 35 (3437)
Semolina 1.99 0.02 1.80 0.01 2.07 0.02 43 (4146) 43 (4146) 1.70 0.03 1.68 0.03 1.69 0.03 43 (4245) 43 (4245) 0.39 0.01 0.36 0.01 0.40 0.01 33 (3235) 33 (3235)
4.37 0.04 4.32 0.04 5.08 0.04 100 100 4.04 0.07 3.69 0.06 4.07 0.07 100 100 1.20 0.01 1.04 0.01 1.20 0.01 100 100
Concentration in mg g1 95% condence interval of the mean. Percent variation versus preceding item: average of the three samples (range). Percent variation versus grain: average of the three samples (range).
at a relatively high concentration in the germ as well, this part accounting for over 10% of the zinc burden in the caryopsis compared to, e.g., 5% of copper (Pieczonka and Rosopulo, 1985). As a consequence, a major proportion of iron, magnesium and zinc in the grain is removed in the conventional roller milling process. Overall, the retention of elements after milling followed the order Se > Ca > Cu > P z K > Fe > Mg z Zn. The sequence Ca > Cu > (P, K, Fe, Mg, Zn) was obtained in previous investigations on durum wheat by Toepfer et al. (1972) (in this case with an inversion between phosphorous and copper) and Brondi et al.
(1984). The same order results from studies on common wheat (Lorenz et al., 1980; Pomeranz and Dikeman, 1983; Toepfer et al., 1972), even though some investigations found a higher proportion of either potassium (Brondi et al., 1984; Lyons et al., 2005; Zhang et al., 1997) or iron (Brondi et al., 1984; Bruggemann and Kumpulainen, 1995; Rao and Deosthale, 1981; Zhang et al., 1997) in our. In most studies the percentage retentions of copper and phosphorous in our were found to be similar (Pomeranz and Dikeman, 1983; Rao and Deosthale, 1981; Toepfer et al., 1972; Zhang et al., 1997), although generally slightly higher for copper. Deviations from the
Table 3 Variations of calcium, iron, zinc, and copper concentration in durum wheat as a consequence of processing and cooking Sample On a fresh weight basis Grain Calcium 1a 2a 3a % vs precb % vs grainc Iron 1a 2a 3a % vs precb % vs grainc Zinc 1a 2a 3a % vs precb % vs grainc Copper 1a 2a 3a % vs precb % vs grainc
a b c
On a dry weight basis Pasta 228 11 248 12 229 11 110 (108111) 63 (6066) 15.7 0.3 13.4 0.2 14.5 0.2 109 (107110) 39 (3740) 13.7 0.4 10.7 0.3 10.8 0.3 104 (103104) 34 (3236) 3.06 0.07 2.71 0.07 3.09 0.07 105 (104107) 54 (5156) Cooked pasta 169 8 193 9 180 9 77 (7479) 48 (4552) 6.3 0.1 5.1 0.1 5.4 0.1 38 (3740) 15 (1515) 5.2 0.2 4.5 0.1 4.4 0.1 40 (3842) 14 (1315) 1.13 0.03 0.98 0.02 1.08 0.03 36 (3537) 20 (1820) Grain 413 20 456 22 383 18 100 100 46.6 0.7 37.4 0.6 41.0 0.7 100 100 43.8 1.4 37.3 1.2 33.0 1.1 100 100 6.15 0.15 5.88 0.14 6.06 0.15 100 100 Semolina 241 12 256 12 235 11 59 (5661) 59 (5661) 16.9 0.3 14.1 0.2 15.1 0.2 37 (3638) 37 (3638) 15.3 0.5 11.8 0.4 11.8 0.4 34 (3236) 34 (3236) 3.35 0.08 3.00 0.07 3.28 0.08 53 (5155) 53 (5155) Pasta 254 12 276 13 254 12 107 (105108) 63 (6166) 17.5 0.3 14.9 0.2 16.1 0.3 106 (104107) 39 (3840) 15.3 0.5 11.9 0.4 12.0 0.4 101 (100101) 34 (3236) 3.41 0.08 3.02 0.07 3.43 0.08 102 (101104) 54 (5157) Cooked pasta 436 21 496 24 462 22 178 (172182) 112 (105121) 16.3 0.3 13.0 0.2 13.9 0.2 89 (8693) 35 (3435) 13.3 0.4 11.6 0.4 11.3 0.4 93 (8797) 32 (3034) 2.92 0.07 2.52 0.06 2.76 0.07 83 (8086) 45 (4347)
Semolina 210 10 224 11 206 10 57 (5460) 57 (5460) 14.7 0.2 12.3 0.2 13.2 0.2 36 (3536) 36 (3536) 13.3 0.4 10.3 0.3 10.3 0.3 33 (3135) 33 (3135) 2.92 0.07 2.62 0.06 2.87 0.07 52 (4953) 52 (4953)
373 18 412 20 346 17 100 100 42.0 0.7 33.7 0.5 37.0 0.6 100 100 39.5 1.3 33.6 1.1 29.8 1.0 100 100 5.54 0.13 5.30 0.13 5.47 0.13 100 100
Concentration in mg g1 95% condence interval of the mean. Percent variation versus preceding item: average of the three samples (range). Percent variation versus grain: average of the three samples (range).
96
Table 4 Variations of selenium concentration in durum wheat as a consequence of processing and cooking Sample On a fresh weight basis Grain 1a 2a 3a % vs precb % vs grainc
a b c
On a dry weight basis Pasta 74.5 1.2 83.5 1.3 87.5 1.4 103 (102103) 83 (7987) Cooked pasta 31.2 0.5 35.8 0.6 36.2 0.6 42 (4143) 35 (3336) Grain 95.2 1.5 109.8 1.8 122.5 2.0 100 100 Semolina 83.5 1.3 93.0 1.5 97.3 1.6 84 (7988) 84 (7988) Pasta 83.1 1.3 93.0 1.5 97.2 1.6 100 (99100) 84 (7987) Cooked pasta 80.3 1.3 92.0 1.5 92.8 1.5 97 (9599) 81 (7684)
Semolina 72.8 1.2 81.2 1.3 85.2 1.4 81 (7785) 81 (7785)
85.8 1.4 99.1 1.6 110.6 1.8 100 100
Concentration in ng g1 95% condence interval of the mean. Percent variation versus preceding item: average of the three samples (range). Percent variation versus grain: average of the three samples (range).
common pattern occasionally observed for potassium, iron, and phosphorous may be due to varying milling conditions among studies, primarily extraction rate (Rao and Deosthale, 1981), and to an uneven release of some elements from milling equipment, especially iron (Cubadda et al., 2005). Furthermore, the proportion of each element in the morphological sections of the wheat kernel is dependent on the genotype and varies among cultivars (Lyons et al., 2005), and this can further explain the slightly different patterns from one study to another. 3.2. Effect of pasta making Overall, pasta making had little effect on element concentrations in parent semolina. Slight, non-statistically signicant enrichments of calcium (107%), iron (106%) and copper (102%) were observed. The concentrations of these elements in the water used for dough preparation were 52 1, 0.10 0.01, and 0.041 0.02 mg L1, respectively, and explained the major part of the calcium increase, a low proportion of that of copper, and 1% of that of iron. Therefore, the release from pieces of equipment used in the pasta making process appeared to be the cause of the concentration increase of iron and, to a lesser extent, of copper. The elemental concentrations found in the commercial pasta samples are summarized in Table 5. These concentrations closely match those of the experimental samples. The only differences were a slightly higher magnesium concentration (p 0.014) and an almost double concentration of selenium (p < 0.001). Lower concentrations of selenium in the experimental samples compared to commercial samples are the consequence of the relatively low levels of plant-available selenium in the areas of Southern Italy from which experimental grain samples originated (Spadoni et al., 2007). The widespread use of wheat imported from selenium-rich areas in industrial semolina production results in a higher level of selenium in commercial pasta. It was investigated as to whether differences in elemental levels between long pasta (spaghetti and related types) and short pasta (macaroni and related types) existed in commercial samples. Differences, indeed, turned out to be negligible, even though a slightly signicantly higher content of copper and zinc in long pasta was detected (0.01 < p < 0.05).
Perhaps of more importance is the variability of concentrations in commercial pasta, which differed widely among minerals. Potassium, magnesium and phosphorous exhibited the lowest CV (56%), followed by copper and iron (9%), calcium (11%), zinc (14%) and selenium (34%). The large spread of selenium levels is not surprising since this element is not essential for plants and its concentration in grains and derived cereal products correlates with plant-available selenium in the soil (Spadoni et al., 2007). The mineral concentrations in pasta in this study are generally similar to those found in recent surveys elsewhere (USDA, 2007), even though more condence can be attributed to the data presented here due to a more representative sampling. However, in earlier studies, substantially lower levels of calcium (Albrecht et al., 1987; Ranhotra et al., 1984, 1985; Yaseen, 1993), copper (Albrecht et al., 1987), potassium (Yaseen, 1993), magnesium (Yaseen, 1993), and phosphorous (Ranhotra et al., 1984; Yaseen, 1993) were reported. It is unclear whether the lower concentrations measured in these earlier studies are the result of analytical shortcomings or reect actual differences due to, e.g., the introduction of new cultivars, modications in agronomic practices, or changes in processing.
3.3. Effect of cooking The effect of cooking can be best assessed considering data expressed on a dry weight basis in Tables 24. Cooking markedly increased calcium concentration of pasta (p < 0.001), which can be ascribed to the calcium content of the cooking water (26 1 mg L1). Except for potassium, the concentrations of the other elements remained unchanged (magnesium) or showed a slight reduction in the range 7997%, which turned out to be signicant only for copper (p < 0.001). Potassium showed a sizeable decrease (p < 0.001), its concentration after cooking being reduced to about 1/4 of that in dry pasta. The results obtained in this study can not be easily compared to those of previous investigations. When expressed on a dry weight basis, the retention values calculated in this study correspond to the apparent retention of nutrients in cooked foods as dened by Murphy et al. (1975). Apparent retentions after pasta cooking were determined in an earlier study, where slightly lower values were found for all minerals (7082%), including calcium, whereas potassium again displayed the highest loss (Yaseen, 1993). However, in this latter study, pasta was cooked in distilled deionized water without addition of salt, which does not allow easy comparison with the results of the present study. A similar approach was used by Ranhotra et al. (1984), but in this case no distilled water was used and calcium turned out to have the highest retention. Other studies (Albrecht et al., 1987; Ranhotra et al., 1985) took into account the loss of solids during cooking and calculated the so-called true retention (Murphy et al., 1975). Albrecht et al. (1987) compared the effect of salt addition when pasta is cooked either in distilled or tap water. Calcium true retention increased from 89% to about 100% when salt was added to distilled water.
Table 5 Element concentrations in commercial pasta samples (N 12)a Cu Fe K Mg P Se Zn Ca (mg g1) (mg g1) (mg g1) (mg g1) (mg g1) (mg g1) (ng g1) (mg g1) Median Mean Min Max CV(%) 266 260 200 314 11 3.30 3.21 2.76 3.56 9 12.9 13.0 11.3 14.7 9 2.07 2.07 1.87 2.29 6 430 430 398 457 5 1.66 1.67 1.51 1.82 6 141 154 84 227 34 12.7 12.9 10.3 15.4 14
a Fresh weight basis (average water content for conversion to dry wt basis is 11.0%).
97
Calcium retention in unsalted tap water was about 129% (average of macaroni and spaghetti) and increased to 158% upon salt addition, with negligible differences following subsequent rinsing. In agreement with the other studies, potassium was found to have the lowest retention and salt addition decreased it, which probably explains the low retention value for this element obtained in the present study. Studies on different cooking approaches may give insight into the effect of each specic medium on mineral retention and suggest explanations for different retention patterns. However, in this study pasta was cooked in the customary way, i.e., in salted tap water, in order to identify the actual mineral content of the nal product when prepared according to common household practice. Using the retention factors obtained in this study, the amount of each mineral provided by a standard serving of pasta (80 g of the uncooked product) containing each mineral at the average concentration found in the commercial samples (Table 5) was calculated. This amount was compared with the Italian daily recommended dietary allowance (RDA) for the adult population (SINU, 1996). It resulted that the proportion of the RDA provided by a daily serving of pasta is on average 22% for selenium, 18% for copper, 10% 14% for zinc (for males and females, respectively), 11% for phosphorous, 9%5% for iron (for males and females, respectively), 4% for calcium, and 1% for potassium. For magnesium, only a range of intakes is established instead of a RDA; if a value one-third above the lower level of this range is chosen as a reference, the proportion of such an amount provided by a daily serving of pasta turns out to be 17%. Overall, pasta appears to be a valuable source of several minerals of importance in human nutrition and well-being. 4. Conclusions This study led to a better understanding of the effect of durum wheat processing on the levels of eight minerals, namely, calcium, copper, iron, magnesium, phosphorous, potassium, selenium, and zinc. For selenium, no data on the changes induced by durum wheat processing were available so far, notwithstanding the ever increasing awareness of the importance of this element to human health. Furthermore, the effect of cooking was investigated in order to determine the retention factors to be used for the estimation of the content of each element in the nal product (as consumed) when the element concentration in the uncooked product is known. Milling was the most important processing step in the production of conventional pasta in regard to the change in content of minerals originally present in the durum wheat grains. At least six groups of elements could be distinguished on the basis of their concentration decrease upon milling. Selenium had the highest retention with concentrations in semolina equal to 77%85% of that in grain (dry weight basis), followed by calcium (54%60%), copper (49%53%), potassium and phosphorous (42%47%), iron (36% 38%), magnesium and zinc (32%36%). Pasta making had little effect on element concentrations in semolina whereas cooking caused negligible to small losses of elements, except for calcium and potassium which greatly increased and decreased their concentration, respectively. Using the retention factors determined by the cooking experiments and the average concentrations ascertained in the commercial pasta samples it was assessed that pasta can provide nutritionally important amounts of several minerals, especially selenium, copper, magnesium, and zinc. Acknowledgements The skilled technical help of Mr. L. Bartoli in the milling of grains and in the pasta making is acknowledged.
References
AACC International, 2000. Approved Methods of the American Association of Cereal Chemists, 10th Ed. The Association, St. Paul, MN, USA. Albrecht, J.A., Asp, E.H., Buzzard, I.M., 1987. Contents and retentions of sodium and other minerals in pasta cooked in unsalted or salted water. Cereal Chemistry 64, 106109. Bock, M.A., 2000. Minor constituents of cereals. In: Kulp, K., Ponte, J.G. (Eds.), Handbook of Cereal Science and Technology, second ed. Marcel Dekker, New York, pp. 479504. Brondi, M., Ciardi, A.M., Cubadda, R., DallAglio, M., Mastino, G.G., Michetti, I., Santaroni, G.P., 1984. Trasferimento di elementi traccia dallambiente alla catena ` alimentare: livelli in grani e loro prodotti. La Rivista della Societa Italiana di Scienza dellAlimentazione 13, 2738. Bruggemann, J., Kumpulainen, J., 1995. The status of trace elements in staple foods. II. Some effects of cereal and potato processing. Zeitschrift fur LebensmittelUntersuchung und-Forschung 201, 711. Carcea, M., Aureli, F., Cubadda, F., 2007. Minerals and trace elements in the Italian wheat and products. Tecnica Molitoria International 58, 129139. Choi, M.K., Kim, E.Y., 2007. Evaluation of dietary manganese intake in Korean men and women over 20 years old. The Journal of the Korean Society of Food Science and Nutrition 36, 447452. Cubadda, F., Baldini, M., Carcea, M., Pasqui, L.A., Raggi, A., Stacchini, P., 2001. Inuence of laboratory homogenization procedures on trace element content of food samples: an ICP-MS study on soft and durum wheat. Food Additives and Contaminants 18, 778787. Cubadda, F., Raggi, A., Testoni, A., Zanasi, F., 2002. Multielemental analysis of food and agricultural matrixes by inductively coupled plasma-mass spectrometry. Journal of AOAC International 85, 113121. Cubadda, F., Raggi, A., Marconi, E., 2005. Effects of processing on ve selected metals in the durum wheat food chain. Microchemical Journal 79, 97102. Guttieri, M.J., Peterson, K.M., Souza, E.J., 2006. Mineral distributions in milling fractions of low phytic acid wheat. Crop Science 46, 26922698. Hattori, H., Ashida, A., Ito, C., Yoshida, M., 2004. Determination of molybdenum in foods and human milk, and an estimate of average molybdenum intake in the Japanese population. Journal of Nutritional Science & Vitaminology 50, 404409. Hussein, L., Bruggemann, J., 1997. Zinc analysis of Egyptian foods and estimated daily intakes among an urban population group. Food Chemistry 58, 391398. Lorenz, K., Loewe, R., Weadon, D., Wolf, W., 1980. Natural Levels of nutrients in commercially milled wheat ours. III. Mineral Analysis. Cereal Chemistry 57, 6569. Lyons, G.H., Genc, Y., Stangoulis, J.C.R., Palmer, L.T., Graham, R.D., 2005. Selenium distribution in wheat grain, and the effect of postharvest processing on wheat selenium content. Biological Trace Element Research 103, 155168. Murphy, E.W., Criner, P.E., Gray, B.C., 1975. Comparisons of methods for calculating retentions of nutrients in cooked foods. Journal of Agricultural and Food Chemistry 23, 11531157. ODell, B.L., de Boland, A., Koirtyohann, S.R., 1972. Distribution of phytate and nutritionally important elements among morphological components of cereal grains. Journal of Agricultural and Food Chemistry 20, 718721. Pieczonka, K., Rosopulo, A., 1985. Distribution of cadmium, copper, and zinc in the caryopsis of wheat (Triticum aestivum L.). Fresenius Journal of Analytical Chemistry 322, 697699. Pomeranz, Y., 1988. In: Pomeranz, Y. (Ed.), Wheat Chemistry and Technology. Chemical composition of kernel structures, Vol. I. American Association of Cereal Chemists, St. Paul, MN, USA, pp. 97158. Pomeranz, Y., Dikeman, E., 1983. Minerals and protein contents in hard red winter wheat ours. Cereal Chemistry 60, 8082. Ranhotra, G.S., Gelroth, J.A., Novak, F.A., Bock, M.A., Winterringer, G.L., Matthews, R.H., 1984. Nutritive value of selected variety breads and pastas. Journal of the American Dietetic Association 84, 322327. Ranhotra, G.S., Gelroth, J.A., Novak, F.A., Bock, M.A., Matthews, R.H., 1985. Retention of selected minerals in enriched pasta products during cooking. Cereal Chemistry 62, 117119. Rao, D.S.S., Deosthale, Y.G., 1981. Mineral and trace element composition of wheat and wheat ours of different extraction rates. Journal of Plant Foods 3, 251257. SINU, 1996. Livelli di Assunzione Raccomandati di Energia e Nutrienti per la Popolazione Italiana (Recommended dietary allowances for energy and nutri` ents for the Italian population). Societa Italiana di Nutrizione Umana, Rome. Spadoni, M., Voltaggio, M., Carcea, M., Coni, E., Raggi, A., Cubadda, F., 2007. Bioaccesible selenium in Italian agricultural soils: comparison of the biogeochemical approach with a regression model based on geochemical and pedoclimatic variables. The Science of the Total Environment 376, 160177. Toepfer, E.W., Polansky, M.M., Heart, J.F., Slover, H.T., Morris, E.R., Hepburn, F.N., Quackenbush, F.W., 1972. Nutrient composition of selected wheats and wheat products. XI. Summary. Cereal Chemistry 49, 173186. UNIPI, 2007. Consumo di pasta alimentare nei diversi paesi (Consumption of alimentary pasta in the different countries). Web site: http://www.unipi-pasta. it/dati/PDF/pdf%202006/TAB31.pdf Tab. 31. USDA, 2007. National Nutrient Database for Standard Reference, Release 20, NBD No. 20420. Web site: http://nal.usda.gov/fnic/foodcomp/cgi-bin/list_nut_edit.pl. Yaseen, A.A.E., 1993. Effect of processing conditions and cooking on retention of minerals in macaroni. Die Nahrung 37, 449455. Zhang, Z.W., Moon, C.S., Watanabe, T., Shimbo, S., Ikeda, M., 1997. Contents of pollutant and nutrient elements in rice and wheat grown on the neighbouring elds. Biological Trace Element Research 57, 3950.

Starch granule size distribution of hard red winter and hard red spring wheat: Its effects on mixing and breadmaking quality
Seok-Ho Park, Jeff D. Wilson*, Bradford W. Seabourn
Grain Marketing and Production Research Center, Agricultural Research Service, Department of Agriculture (USDA), 1515 College Avenue, Manhattan, KS 66502, USA
Article history: Received 24 April 2008 Received in revised form 2 July 2008 Accepted 10 July 2008 Keywords: Starch granule size distribution HRW HRS Mixing property Breadmaking quality Crumb grain score Optimum range of B-granules
a b s t r a c t
Starch was isolated from 98 hard red winter (HRW) wheat and 99 hard red spring (HRS) wheats. Granule size/volume distributions of the isolated starches were analyzed using a laser diffraction particle size analyzer. There were signicant differences in the size distribution between HRW and HRS wheats. The B-granules (<10 mm in diameter) occupied volumes in the range 28.549.1% (mean, 39.9%) for HRW wheat, while HRS wheat B-granules occupied volumes in the range 37.156.2% (mean, 47.3%). The mean granule sizes of the distribution peaks less than 10 mm in diameter also showed a signicant difference (HRW, 4.32 vs. HRS, 4.49 mm), but the mean sizes of the distribution peaks larger than 10 mm were not signicantly different (21.54 vs. 21.47 mm). Numerous wheat and our quality traits also showed signicant correlation to starch granule size distributions. Most notably, protein content was inversely correlated with parameters of B-granules. Crumb grain score appeared to be affected by starch granule size distribution, showing signicant inverse correlations with B-granules. Furthermore, the linear correlations were improved when the ratio of B-granules to protein content was used, and the polynomial relation was applied. There also appeared to be an optimum range of B-granules for different protein content our to produce bread with better crumb grain. Published by Elsevier Ltd.
1. Introduction Starch is an important part of wheat endosperm, not only because starch accounts for 6573% of dry our mass when milling extraction is <80% (Pomeranz, 1988), but also because wheat starch has unique properties in breadmaking that are not replaceable by other starches from corn, potato, and cassava (Hoseney et al., 1971; Sahlstrom et al., 1998; Sollars and Rubenthaler, 1971). Wheat starch granules have been reported to have bimodal size distribution (Dengate and Meredith, 1984; Evers and Lindley, 1977; Karlsson et al., 1983; Morrison and Scott, 1986; Simmonds and OBrien, 1981; Stoddard, 1999), but trimodal size distributions have also been reported (Bechtel et al., 1990; Raeker et al., 1998). The large A-granules (generally larger than 10 mm in diameter) are formed rst in developing endosperm, whereas the small B-granules (smaller than 10 mm in diameter) are formed late in kernel development (Dengate and Meredith, 1984; Karlsson et al., 1983; Sandstedt, 1961). Bechtel et al. (1990) reported the formation of very small C-type granules (less than 5 mm) that were initiated very late in grain lling. Different size starch granules have different physical, chemical, and functional properties (Chiotelli and Le
Meste, 2002; Dronzek et al., 1972; Eliasson, 1989; Kulp, 1973; Lineback, 1984; Meredith, 1981; Morrison and Gadan, 1987; Nikuni, 1978; Park et al., 2004; Peng et al., 1999; Soulaka and Morrison, 1985a). In addition, there has been considerable debate on the role of B-granules in breadmaking, which was discussed in a previous report by Park et al. (2005). Starch granule size distribution is an important factor that affects the quality of many nal products. A wide variation (1750%) of B-starch granule volume was found from a survey of hexaploid wheats, suggesting prospects of genetic manipulation of granule size distribution (Stoddard, 1999). However, there has been limited research conducted to nd relationships between starch granule size distribution and nal product quality. Also, there are no reports comparing the starch granule size distribution of HRW and HRS. Therefore, the objective of this study was to investigate starch granule size distribution of HRW and HRS, and determine its relationship to wheat, our, dough mixing, and breadmaking properties. 2. Experimental 2.1. Materials
* Corresponding author. Tel.: 1 785 776 2763; fax: 1 785 537 5534. E-mail address: jeff.d.wilson@usda.ars.gov (J.D. Wilson). 0733-5210/$ see front matter Published by Elsevier Ltd. doi:10.1016/j.jcs.2008.07.011
A total of 197 wheat samples, including 98 hard red winter and 99 hard red spring harvested in 2002 and 2003, were provided by
S.-H. Park et al. / Journal of Cereal Science 49 (2009) 98105
99
Nomenclature A-granules B-granules HRS HRW r larger than 10 mm in diameter smaller than 10 mm in diameter hard red spring hard red winter simple correlation coefcient
R2 * ** *** SEM SKCS
coefcient of determinant P < 0.01 P < 0.001 P < 0.0001 standard error of the mean single kernel characterization system
the Federal Grain Inspection Service (FGIS) Technical Center, (Kansas City, MO), Grain Inspection, Packers, and Stockyards Administration (GIPSA), U.S. Department of Agriculture. Detailed information about the samples has been previously reported (Maghirang et al., 2006). 2.2. Starch isolation and determination of granule size distribution Starch was isolated by enzymatic digestion using pepsin A (P7012, Sigma, St. Louis, MO), hemicellulase 90 (90,000 U/g activity, Amano Enzyme U.S.A., Lombard, IL), and cleaned further by a detergent mix (5% SDS, 5% Triton X-100, 5% Tween 40, and 5% Triton X-15) (Bechtel and Wilson, 2000). The size distribution of isolated starch granules was measured using a single wavelength Beckman Coulter LS 13 320 Particle Size Analyzer (Beckman Coulter, Miami, FL) with the Universal Liquid Module for liquid-based measurements. Each starch sample was slurried with 1.0 mL of water and vortexed before analysis. The standard refractive indices used were 1.31 for water and 1.52 for starch, which is within the sample concentration range of the instruments specications. Volumes of all starch granules were calculated on the assumption that all granules were spherical in shape. 2.3. Evaluation of wheat and our properties The following properties of wheat were analyzed: test weight (lb/bu) by American Association of Cereal Chemists (AACC, 2000) Approved Method 55-10; protein content using near-infrared reectance (AACC Approved Method 39-25); single kernel hardness (AACC Approved Method 55-31), weight, and size using the SKCS 4100 (Perten, Springeld, IL); and ash content (AACC Approved Method 08-01). Wheat was milled using a Brabender Quadrumat Sr. experimental mill (AACC Approved Method 26-10A). Flour protein and ash content were measured using AACC Approved Method 39-11 and 08-01, respectively. Flour color (L*, a*, and b*) was determined using a colorimeter (CR-300, Minolta, Osaka, Japan). Mixing characteristics of our were evaluated using Mixograph (AACC Approved Method 54-40A) and Farinograph (AACC Approved Method 54-21). A modied optimized straightdough breadmaking method (AACC Approved Method 10-10B) was used for evaluation of experimental breadmaking properties of ours. The detailed baking method and description of crumb grain score were reported previously (Park et al., 2004). All tests were conducted with at least 2 duplicates. 2.4. Statistical analysis A complete randomized experimental design was used. The difference between starch granule size and volume distributions of HRW and HRS was analyzed using the General Linear Models procedure of the Statistical Analysis System (SAS Institute, Cary, NC). Statistical abbreviations were simple correlation coefcient (r), coefcient of determinant (R2), P < 0.01 (*), P < 0.001 (**), P < 0.0001 (***), and standard error of the mean (SEM).
3. Results and discussion 3.1. Starch granule size distributions of HRW and HRS Fig. 1 shows a typical bimodal size distribution of wheat starch granules, plotting proportions by volume % of granules in equal diameter intervals against diameter. Point 1 and 3 represent the most frequent differential volume (%) of the small and large granules, and Point 2 is differential volume (%) between small and large granules that sets apart small and large starch granules. The respective integrating areas (%) (A, B, and C) represent proportion of volume distribution (%) split by each Point. Starch granule size distributions of HRW (n 98) and HRS (n 99) showed signicant differences in mean values of different aspects of granule size distribution including differential volume, diameter, and volume distribution (Table 1). Table 1 shows that HRW has less B-granules as indicated by lower differential volume compared to HRS wheat (2.25% vs. 2.80%, respectively). Differential volumes of HRW and HRS wheats at Point 2, however, were not signicantly different. At Point 3, differential volume for HRW wheat was higher (6.55%) than that of HRS (6.03%.), and is due to the fact that differential volume (%) represents proportionality. HRW and HRS wheats have similar starch granule size distribution range from less than 1 mm to about 40 mm, thus, a lower % in one proportion results in higher % proportion elsewhere. Fig. 2 shows typical starch granule size distribution curves of HRW and HRS wheats, demonstrating that HRS wheat had higher and lower differential volumes at Point 1 and 3, respectively, compared with HRW wheat. It should be noted that even though these curves (differential volume) were obtained from calculations using real numbers of starch granules at a specic size, the differential volume represents the proportion of starch granule size distribution in the kernel and not the actual volume. The mean values of starch granule diameter at each point, compared to differential volume, represent actual size. HRW had signicantly smaller size of B-granules (4.32 mm) compared with
8 7
Point 3
Differential volume (%)
6 5 4 3 2 1 0 0 1 2 6 16 40
Point 1 Point 2
Size (diameter, m)
Fig. 1. Starch granule size distribution indicating high and low points (Point 1, 2, and 3) of differential volume and their volume percents (Area A, B, C).
100
Table 1 Starch granule size distribution in differential volume (%), diameter (mm), and volume percent distribution (%) of HRW and HRSa HRW Size Distributionb Differential volume (%) Point 1 Point 2 Point 3 Diameter (mm) Point 1 Point 2 Point 3 Volume distribution (%) A AB A BC
a b c
HRS Minimum 1.30 0.91 5.17 Maximum 3.15 1.96 7.91 SEMc 0.025 0.015 0.029 Mean 2.80*** 1.56 6.03*** Minimum 1.91 0.73 4.74 Maximum 4.12 2.14 7.20 SEMc 0.029 0.022 0.034
Mean 2.25*** 1.53 6.55***
4.32*** 8.72*** 21.49
3.52 5.61 18.86
5.11 10.78 24.95
0.030 0.063 0.087
4.49*** 9.46*** 21.46
3.86 8.15 18.86
5.11 10.78 24.95
0.029 0.043 0.088
23.53*** 38.01*** 75.78***
14.94 12.57 57.59
30.87 49.07 89.00
0.245 0.402 0.300
29.35*** 46.88*** 80.05***
20.06 34.03 73.34
38.24 57.81 87.30
0.254 0.308 0.180
*** Mean values of HRW and HRS in the same row are signicantly different at P 0.0001. Explanation of specic points and area are given in Fig. 1. SEM standard error mean.
HRS (4.49 mm) at Point 1 (Table 1). The observation that HRW wheat had a smaller size of B-granules compared with HRS wheat was conrmed by the fact that the threshold diameters at Point 2 were 8.72 mm for HRW wheat and 9.46 mm for HRS wheat. The mean diameter of A-granules, however, was not signicantly different at Point 3 (21.49 and 21.46 mm, respectively). The HRW and HRS wheats showed a wide range of volume distributions in terms of A- and B-granules. The volume distribution of HRW wheat B-granules ranged from 12.57% to 49.07%, while HRS wheat B-granules ranged from 34.03% to 57.81% (Table 1, Fig. 1). The mean volume distributions of HRW wheat B-granules were signicantly lower (23.53% and 38.01% for area A and A B, respectively) than that of HRS (29.35% and 46.88% for area A and A B, respectively). It is obvious that HRW wheat had smaller B-granules in size and number. The volume distributions of subdivided ranges of starch granule size distribution are shown in Table 2. This data conrms that HRW wheat has signicantly smaller proportions of B-granules (less than 10 mm) compared with HRS wheat in our study. Specic volume ranges for A- and B-granules derived from HRW and HRS wheats have not been reported in the literature, and most
7
HRS
HRW
0 0 1 2 6 16 40
Size (diameter, m)
Fig. 2. Starch granule size distribution of HRW (GIPSA identication number: 03027680) and HRS (GIPSA identication number: 03036243).
previous reports on the proportion of B-granules in total starch were on a weight basis (Brocklehurst and Evers, 1977; DAppolonia and Gilles, 1971; Dengate and Meredith, 1984; Evers, 1973; Evers and Lindley, 1977; Hughes and Briarty, 1976; Meredith, 1981; Park et al., 2004; Soulaka and Morrison, 1985a). Our data, however, can still be compared with previous reports because the difference in density of A- and B-starch granules is negligible (Dengate et al., 1979; Meredith, 1981). Our volume data on the range of B-granules was higher than most previous reports. DAppolonia and Gilles (1971) reported less than 10% B-granules by weight in total starch from 12 HRS ours, whereas others reported 30% B-granules (Evers, 1973; Evers and Lindley, 1977; Hughes and Briarty, 1976). Park et al. (2004) observed 15.727.0% B-granules from 12 HRW wheat ours, and Soulaka and Morrison (1985a) found 1335%. Stoddard (1999) reported a similar range of B-granule volume (%), showing 1750% of B-granules using laser diffraction sizing methods from 130 Australian hexaploid wheat cultivars and Landraces from Asia. Bechtel et al. (1990) found 48% of B- and C-type granules (less than 15.9 mm) from mature HRW wheat, but this discrimination was larger than the commonly used 10-mm range. These different observations on the proportion of B-granules in total starch were most likely due to different cultivars, growing conditions, and starch extraction and analysis methods. The smaller size and number of B-granules in HRW wheat compared with HRS wheat could be a characteristic of wheat class, or it could be caused by higher temperature during the grain-ling period. Dengate and Meredith (1984) found that drought affected granule size distribution (weight %), decreasing B-granules (410 mm) and increasing smaller A-granules (1020 mm). The volume % of B-granules decreased when growth temperature was increased from 15 C to 40 C during grain-lling period (Shi et al., 1994). High temperature also seemed to be associated with a reduction in the number of B-granules (Bhullar and Jenner, 1985). Barley, which also has a bimodal starch granule size distribution, showed a similar low ratio in number of B-granules induced by high temperature (MacLeod and Duffus, 1988; Tester et al., 1991). These environmental effects could cause a difference in the grain-lling pattern between A- and B-granules. The A-granule starts to form in the amyloplast at about 45 days after anthesis (Bechtel et al., 1990; Parker, 1985), and continues to increase in size until reaching a maximum diameter of 2550 mm at physiological maturity (Bechtel et al., 1990; Dengate and Meredith, 1984; Simmonds and OBrien, 1981). The nal number of A-amyloplasts, however, is achieved much earlier at approximately seven days post-anthesis when cell division ceases (Briarty et al., 1979). On the other hand,
Differential volume (%)
S.-H. Park et al. / Journal of Cereal Science 49 (2009) 98105 Table 2 Volume percent distribution (%) on different ranges of granule size of HRW and HRSa HRW Range (mm) <2 25 510 1020 2030 >30 <5 <10
a b
101
HRS Minimum 3.9 12.0 11.8 19.5 18.6 2.2 16.4 28.5 Maximum 8.2 23.1 18.5 36.7 27.7 22.9 30.6 49.1 SEMb 0.06 0.16 0.10 0.20 0.12 0.24 0.20 0.28 Mean 8.2*** 22.1*** 17.1*** 24.6*** 21.6*** 6.5*** 30.2*** 47.3*** Minimum 6.1 16.5 13.7 17.7 16.3 3.3 23.2 37.1 Maximum 10.1 26.8 22.1 29.9 26.1 13.3 36.6 56.2 SEMb 0.05 0.13 0.11 0.18 0.13 0.13 0.17 0.25
Mean 6.6*** 18.5*** 14.7*** 26.8*** 23.7*** 9.6*** 25.2*** 39.9***
*** Mean values of HRW and HRS in the same row are signicantly different at P < 0.0001. SEM standard error mean.
B-granules are initiated at 1012 days after anthesis and continue to enlarge until 21 days post-anthesis, and up to approximately 9 mm by maturity (35 days after anthesis) (Bechtel et al., 1990). Thus, considering the nature of the growth pattern of B-granules that initiate and enlarge toward the end of the grain-lling period, a high temperature event could cause a decrease in starch synthase activity and reduce the duration of grain lling, resulting in smaller size and number of B-granules in the endosperm. The U.S. Wheat Associates harvest report of 2002 stated that 2002 HRW wheat production was the lowest since 1970 because of persistent drought conditions throughout the Great Plains region (U.S. Wheat Associate, 2002). In this study, however, we could not verify that drought conditions or high temperatures caused smaller size and number of B-granules in HRW wheat due to lack of detailed weather information concerning growing conditions for tested wheats, and no HRW wheats were grown in the HRS wheat growing region or vice versa. In addition, there is no previous comprehensive study on this subject because various U.S. wheat varieties adapt to diverse growing regions and respond differently to the environment (Altenbach et al., 2003). It must also be pointed out that our volume data, which was generated by laser diffraction analysis, was calculated based on the assumption that all starch granules were spherical in shape. Starch granules larger than 5 mm in diameter are typically oblate spheroids in shape (Bechtel et al., 1990; Bechtel and Wilson, 2000). Therefore, differential volume of granules over 5 mm in diameter in this study could be over-estimated. In particular, the volume of small granules between 5 and 10 mm could be over-estimated because the number of B-granules composes over 90% of total number of starch granules (Evers and Lindley, 1977; Morrison and Gadan, 1987; Park et al., 2004). Wilson et al. (2006) found that volume data obtained from laser diffraction analysis needed a correction factor for better correlation with data from digital image analysis. 3.2. Relationships with wheat and our properties Starch granule size distribution showed many signicant correlations with wheat and our properties. HRS wheat, in particular, was highly correlated when compared with HRW wheat (Table 3). Since we do not know the cause of this difference, our discussion at this point will focus primarily on HRS wheat. Protein content in HRS wheat showed negative correlations with differential volume at Point 1 (r 0.72***), but displayed positive relationships at Point 2 (r 0.48***) and 3 (r 0.62***). HRW showed similar relationships, but there was no signicant correlation at Point 3. Starch granule size (diameter) at Points 1 and 2 was negatively correlated with HRS wheat protein content (r 0.52*** and r 0.67***, respectively), although there was no signicant correlation at Point 3. In addition, area range A and A B, representing volume proportion of B-granules, showed negative correlations with wheat protein content. All relationships
between wheat protein content and starch granule size distribution indicated that size and number of B-granules decrease when protein content increases, or vice versa. Altenbach et al. (2003) reported that in HRS wheat exposed to high temperature regimes (37/17 or 37/27 C day/night), onset and cessation of starch accumulation occurred earlier and overall accumulation period was shortened. They, however, observed slight changes in the timing and duration of protein accumulation during the temperature regime. In addition, previous studies reported that high temperature stress during maturation signicantly reduced kernel weight as much as 85% (Stone and Nicolas, 1994; Tashiro and Wardlaw, 1990). Therefore, it is highly likely for individual kernels to have high protein content on a percentage basis when they were exposed to high temperature, and as described earlier, high temperature affects the starch granule synthase, thus decreasing the number and size of B-granules. This may be an explanation as to why wheat protein content (%) showed inverse correlations to B-granules, including differential volume at Point 1 (r 0.72***), diameter at Points 1 and 2 (r 0.52*** and 0.67***, respectively), and area range of A and A B (r 0.68*** and 0.66***, respectively). On the other hand, test weight was positively correlated with B-granule parameters, such as differential volume at Point 1 (r 0.72***), diameter at Points 1 and 2 (r 0.45*** and 0.64***, respectively), and area range A and A B (r 0.70*** and 0.71***, respectively). Test weight generally increased when kernel weight increased (Ohm et al., 1998), and kernel weight and dimensions are decreased by high temperature during grain lling (Gibson et al., 1998; Gibson and Paulsen, 1999; Tashiro and Wardlaw, 1990). Since wheat kernels exposed to high temperature tend to have a low proportion of B-granules, this could explain the positive correlation between test weight and volume of B-granules. The proportion of large kernels and single kernel weight and size showed similar relationships with parameters of starch granule size distribution because of the reasons we mentioned earlier (Table 3). Therefore, our data conrms previous reports and links the separate observations with this large data set. Flour protein showed similar relationships as wheat protein, as expected. Flour brightness (L), showed positive correlation with parameters of B-granules, possibly due to the inverse correlation between our brightness and our protein content (r 0.56***, data not shown). In other words, our color gets darker as protein content increases, and as explained previously, wheat kernels with high protein content tend to have a smaller ratio of B-granules. 3.3. Relationships with mixing properties The relationship between mixing properties and starch granule size distribution showed similar trends to the relationship between protein content and starch granule size distribution (Table 4). In HRS wheat, differential volume at Point 1 (r 0.67***), diameter
102
Table 3 Correlations between starch granule size distribution of HRW and HRS and wheat and our properties Differential volume (%) Point 1 HRW Wheat protein (%) Test weight (bu/lb) 0.39 *** Point 2 0.40 *** 0.36 ** Point 3 Diameter (mm) Point 1 Point 2 0.49 0.32 *** * Point 3 0.47 *** Area range (%) A AB ABC
Kernel size distribution (%) Large kernel 0.56 Medium kernel 0.56 Small kernel 0.28 Single kernel character Weight (mg) 0.40 Size (mm) 0.49 Hardness 0.29 Flour protein (%) 0.35 Flour color: L 0.32 HRS Wheat protein (%) Test weight (bu/lb)
*** *** *
0.32 0.33
* **
0.40 0.39
*** ***
0.33 0.31 0.28
** * *
0.61 0.60 0.36
*** *** **
0.28 0.28
* *
0.40 0.39 0.27
*** *** *
0.41 0.41
*** ***
*** *** * ** *
0.38 0.45 0.30 0.41 *** 0.41
*** *** * ***
0.30
0.48 0.53 0.31 0.44 0.32
*** *** * *** *
0.34 0.30 0.45
** * ***
0.33
**
0.29 0.36
* **
0.72 0.72
*** ***
0.48 0.41
*** ***
0.62 0.67
*** ***
0.52 0.45
*** ***
0.67 0.64
*** ***
0.68 0.70
*** ***
0.66 0.71
*** ***
0.30 0.26
* *
Kernel size distribution (%) Large kernel 0.68 Medium kernel 0.68 Small kernel 0.29 Single kernel character Weight (mg) 0.59 Size (mm) 0.63 Hardness Flour protein (%) 0.72 Flour color: L 0.53
*** *** *
0.50 0.50 0.28
*** *** *
0.60 0.60 0.29
*** *** *
0.53 0.53
*** ***
0.70 0.70 0.28
*** *** *
0.30 0.30
* *
0.67 0.67
*** ***
0.63 0.64
*** ***
0.29 0.29
* *
*** *** *** ***
0.49 0.44 0.51 0.33
*** *** *** **
0.53 0.58 0.59 0.50
*** *** *** ***
0.56 0.57 0.29 0.54 0.44
*** *** * *** ***
0.48 0.54 0.70 0.52
*** *** *** ***
0.27 0.27 0.29
* * *
0.59 0.63 0.68 0.54
*** *** *** ***
0.45 0.53 0.64 0.50
*** *** *** *** 0.28 0.31 * *
*, **, and *** signicant correlations at P < 0.01, P < 0.001, and P < 0.0001, respectively.
at Points 1 and 2 (r 0.53***, 0.65***, respectively), and area range A and A B (r 0.64***, 0.59***, respectively) showed signicant inverse correlations with mixing absorption. This observation is most likely due to mixing absorption being generally positively correlated to protein content (Ohm and Chung, 1999; Park et al., 2006). The B-granules were reported to have higher
Table 4 Correlations between starch granule size distribution and mixing properties Differential volume (%) Point 1 HRW Mixograph Absorption Mix time Mix tolerance Farinograph Absorption Development time Stability Tolerance Breakdown HRS Mixograph Absorption Mix time Mix tolerance Farinograph Absorption Development time Stability Tolerance Breakdown Point 2 Point 3 Diameter (mm) Point 1
swelling power (intragranular plus entrained intergranular water), which was due to higher water absorption capability, and was triggered only when temperature was much higher than normal mixing temperature, above 90 C (Kulp, 1973; Park et al., 2004; Seib, 1994; Wong and Lelievre, 1982). Haraszi et al. (2004) observed a decrease in water absorption when protein content decreased
Area range (%) Point 2 Point 3 A AB ABC
0.30
0.42
*** 0.26 *
0.40
***
0.42
***
0.32 0.34 0.29
* ** *
0.35 0.32
** *
0.28
0.30 0.31
* *
0.36 0.39
** ***
0.36
**
0.34
**
0.67 0.59 0.55
*** *** ***
0.47 0.29 0.35
*** * **
0.56 0.53 0.47
*** *** ***
0.53 0.36 0.41
*** ** ***
0.65 0.45 0.39
*** *** ***
0.30
0.64 0.58 0.55
*** *** ***
0.59 0.57 0.49
*** *** ***
0.27 0.35 0.32
* ** *
0.29 0.59 0.48 0.33 0.62
* *** *** ** ***
0.33 0.38 0.37
** *** **
0.49 0.41 0.52
*** *** ***
0.37 0.46 0.43
** *** ***
0.38 0.54 0.48 0.39 0.60
*** *** *** *** ***
0.34
**
0.29 0.59 0.48 0.27 0.61
* *** *** * ***
0.35 0.61 0.40 0.29 0.61
** *** *** ** ***
0.31
0.29
103
with starch addition. Also, mixing absorption is highly inuenced by protein subclass, 50% 1-propanol insoluble polymeric protein content in our (Park et al., 2006). HRW wheat showed similar trends, but with rather weak and insignicant relationships. Mixograph mixing time in HRS wheat showed negative correlations with B-granules, whereas in HRW wheat, a positive correlation was observed between mix time and differential volume at Point 1 (Table 4). Park et al. (2004) found that mix time did not show signicant correlations to protein content with 49 HRW wheat ours due to contrasting effects among protein subclass on mix time. However, they suggested that mix time would be shorter as protein content increases because they also found that 50% 1-propanol insoluble polymeric protein based on total protein was positively correlated with mix time and tended to decrease with increasing total protein content. Consequently, considering significant inverse correlations between protein content and parameters of B-granules, the relationship between mix time and parameters of B-granules was expected to be positive. The positive correlation (r 0.33**), however, was obtained only from HRW wheat. The inverse correlations between HRS wheat and parameters of B-granules were probably due in part to a positive correlation between protein content and mix time (r 0.44***, data not shown) in HRS wheat. In HRW wheat, there was no signicant correlation between protein content and mix time. It has also been reported that B-granules require shorter optimum mix time compared with A-granules (DAppolonia and Gilles, 1971; Petrofsky and Hoseney, 1995). Chiotelli and Le Meste (2002) also reported that B-granules had a higher afnity for water than the A-granules at room temperature, resulting in faster hydration. Therefore, the present work suggests that a higher proportion of B-granules in HRS wheat may partially account for the inverse relationship between mix time and parameters of B-granules. Farinograph absorption showed weak but signicant inverse correlations to parameters of B-granules in HRS wheat, whereas in HRW wheat, diameter at Point 2 showed signicant inverse correlation (Table 4). A similar reason, as given for Mixograph mix absorption, could explain these relationships. Farinograph absorption was positively correlated with wheat protein (r 0.71***, data not shown), and parameters of B-granules were inversely correlated to protein content, resulting in negative correlations between Farinograph absorption and parameters of B-granules. Soh et al. (2006) observed a signicant increase in Farinograph absorption at constant protein content (17.417.7%) as % B-granules increased from 17% to 32.4%. Therefore, it appears that starch granule size distribution may affect Farinograph absorption when protein content and quality are invariable. Our data suggests that starch granule size distribution may not be an important variable for mix absorption when protein content varies. The protein content in the HRS wheat our varied widely from 10.6% to 17.8% (data not shown), and the protein composition (quality) was altered as the protein content increased (Park et al., 2006). The relationship between Farinograph development time and starch granule size distribution was similar to the relationship between Mixograph mix time and starch granule size distribution. The same explanation would be applicable for both relationships. Farinograph stability in HRS wheat also showed inverse correlation with parameters of B-granules. Several authors have reported the effects of starch on rheological properties. B- and nonwheat starch granules have been reported to result in large rheological differences using a constant vital gluten ratio in the reconstituted dough (Petrofsky and Hoseney, 1995). They suggested that there were strong interactions between vital gluten and nonwheat starches and wheat B-granules, resulting in less extensibility. Miller and Hoseney (1999) found that starch isolated from Kansas grown wheat gave signicantly lower elastic modulus (G0 ) and viscous modulus (G00 ) than starches from other Kansas and strong
Canadian wheats when dough was prepared from the same gluten source. Thus, it appears that starch from different cultivars and different size distributions makes a difference in determining the rheological properties of dough when a constant gluten source is used. The inverse relationships between Farinograph stability and parameters of B-granules, however, are hard to explain from previous reports (Miller and Hoseney, 1999; Petrofsky and Hoseney, 1995) for two reasons. First, the present study used actual HRS wheat ours, and not blends of starches and constant gluten, as did previous researches, which had various components that could affect mixing properties. Second, there is limited research to connect rheological parameters of B-granules to Farinograph stability. It is generally recognized that rheological development in dough is triggered by the formation of a continuous gluten network (MacRitchie, 1992), with starch granules embedded in the network. In addition, previous studies reported that breakdown (loss of stability) during over-mixing was caused by depolymerization of the protein network (Danno and Hoseney, 1982; Skerritt et al., 1999; Tanaka and Bushuk, 1973; Weegels et al., 1996). Therefore, with large variation in our quality including protein content (10.6 17.8%, 14% moist base), Farinograph absorption (59.873.1%), development time (5.244.5 min), and loaf volume (8031238 cm3) (data not shown), inverse relationships between Farinograph stability and parameters of B-granules could be explained by the positive correlations between protein content and Farinograph stability (r 0.46***, data not shown). As previously discussed, protein content was inversely correlated to parameters of B-granules, possibly giving the inverse correlation between Farinograph stability and parameters of B-granules. Farinograph breakdown showed a similar relationship with stability, and a similar explanation could be applicable. Protein content and breakdown were highly positively correlated (r 0.79***, data not shown). The HRW wheat showed a similar trend in these relationships, but was weaker with many non-signicant correlations. 3.4. Relationships with breadmaking parameters The breadmaking parameters showed signicant correlations to starch granule size distribution (Table 5). A greater number of the parameters of HRS wheat, again, were signicantly correlated. The trends in relationships were similar for HRW and HRS wheat, so our discussion will focus on HRS wheat. Baking absorption and mix time were positively correlated to Mixograph absorption (r 0.85***), mix time (r 0.93***), Farinograph absorption (r 0.73***) and development time (r 0.77***) (data not shown). Loaf volume was inversely correlated with parameters of B-granules. Differential volume at Point 1 showed the greatest inverse correlation (r 0.68***), followed by area range A (r 0.66***), A B (r 0.62***) and diameter at Point 1 (r 0.52***). Park et al. (2005) summarized previous studies using four different concepts concerning the effects of small starch granule size on breadmaking properties: benecial (Hayman et al., 1998; Sahlstrom et al., 1998; Van Vliet et al., 1992); detrimental(DAppolonia and Gilles, 1971; Kulp, 1973); little effect (Hoseney et al., 1971); and optimum ratio of A- and B-granules (Lelievre et al., 1987; Park et al., 2004; Soulaka and Morrison, 1985b). More recently, Park et al. (2005) conrmed their previous observation that there seemed to be an optimum weight ratio of Aand B-granules for crumb grain score, and there was no benet of small starch granules to loaf volume and crumb grain score when used with a constant source of gluten. Considering previous controversial results, it is difcult to state that B-granules affect loaf volume negatively. It is more appropriate to admit that the relationships were obtained due to a positive correlation between protein content and loaf volume (r 0.91***, data not shown), and
104
Table 5 Correlations between starch granule size distribution and straight-dough breadmaking parameters Differential volume (%) Point 1 HRW Breadmaking Absorption Mix time Proof height Crumb grain score Loaf volume HRS Breadmaking Absorption Mix time Proof height Crumb grain score Loaf volume Point 2 Point 3 Diameter (mm) Point 1 Point 2 Point 3 Area range (%) A AB ABC
0.35 0.46 0.35 *** ** 0.38 0.35 0.43
** *** ** *** 0.31 * 0.45 0.28 *** *
0.28 0.51 0.42
* *** ***
0.26 0.26 0.40
* * *** 0.28 * 0.27 *
0.57 0.52 0.56 0.28 0.68
*** *** *** * ***
0.30 0.34 0.45
* ** ***
0.52 0.46 0.51 0.27 0.60
*** *** *** * ***
0.40 0.34 0.37 0.41 0.52
*** ** ** *** ***
0.57 0.43 0.47 0.67
*** *** *** ***
0.28
0.28
0.56 0.52 0.54 0.27 0.66
*** *** *** * ***
0.58 0.53 0.52 0.62
*** *** *** ***
0.37 0.36
** **
0.28
*, **, and *** signicant correlations at P < 0.01, P < 0.001, and P < 0.0001, respectively. All correlations are signicant in this table at P 0.01 (>r j0.261j), P 0.001 (>r j0.326j), and P 0.0001 (>r j0.374j).
protein content was inversely correlated with parameters of B-granules. Crumb grain score showed a generally weak but signicant inverse linear correlation to differential volume at Point 1 (r 0.28*), diameter at Point 1 (r 0.41***), and area range of A (r 0.27*). These inverse relationships agreed with results from Park et al. (2005) where the authors found the lowest value of crumb grain score and neness from the bread baked with 100% B-granules. In this study, crumb grain score was not signicantly correlated to protein, consequently relationships were independent of protein content. It should be pointed out that there were no other signicant correlations to crumb grain score from 68 wheat and our quality parameters, except wheat kernel weight and size (r 0.36***, respectively, data not shown), which were positively correlated to parameters of B-granules. So, even though those correlation values were low, we believe it has signicance. 3.5. Polynomial relationships between B-granules and crumb grain score Several authors reported that there seemed to be an optimum weight % ratio of A- and B-granules for breadmaking (Lelievre et al., 1987; Park et al., 2004, 2005; Soulaka and Morrison, 1985b) and spaghetti production (Soh et al., 2006). A polynomial relationship was applied between B-granules and crumb grain score, resulting in improved correlations for both HRW and HRS wheats (Table 6). Correlation values of HRS wheat increased from 0.073* to 0.154*** for area A, and 0.013 to 0.088* for area A B. In addition, we found that the linear relationships improved when the ratio of volume % of B-granules to our protein content was used. The correlation
Table 6 Linear and polynomial correlations of B-type granule volume % and ratio with protein contents to crumb grain score Crumb grain score vs. Linear correlation R2 0.004 0.005 0.130 0.136 0.073 0.013 0.089 0.049 Polynomial correlation R2 0.018 0.039 0.141 0.158 0.154 0.088 0.222 0.164
improvements were more obvious for HRW wheat, which improved from 0.004 to 0.130** for area range A, and from 0.005 to 0.136** for area A B. The linear correlations between volume % of B-granules and crumb grain score were improved from 0.004 to 0.141*** and from 0.005 to 0.158*** for HRW wheat, and from 0.073* to 0.222*** and from 0.013 to 0.164*** for HRS wheat after obtaining the ratios with protein content and applying polynomial correlation. Therefore, it appears that there is an optimum range of volume % B-granules, and the optimum range could vary depending on protein content. Lelievre et al. (1987) found different optimum starch size fractions for different protein concentrations when used for breadmaking. Park et al. (2005) suggested that high water absorption during baking and/or high surface area of B-granules could be responsible for gas cell stabilization and the resultant crumb grain score. It has been shown that proteinstarch interactions produce different rheological properties depending on variety and granule size (Miller and Hoseney, 1999; Petrofsky and Hoseney, 1995). Also, bulk rheological properties of dough could affect overall gas cell stability (Van Vliet et al., 1992). Consequently, the reason for different optimum weight % of B-starch granules for different protein contents may be due to rheological properties imparted by proteinstarch interactions. The higher water absorbing properties of B-granules could pull water from the attached protein matrix and liquid lm in the dough system during baking. Gan et al. (1995) proposed that gas cells are stabilized by a continuous liquid lm on the proteinstarch matrix. Depending on how extreme the situations are, e.g. low protein content with high content of B-granules (too stiff) vs. high protein content with low content of B-granules (too viscous), the overall viscoelastic properties and gas cell stability could be changed, resulting in different crumb structures.
4. Conclusion Starch granule size distributions of HRW and HRS wheats showed signicant differences in differential volume, diameter, and volume distribution. HRW has smaller size and proportion of B-granules than HRS wheat. The reason is not clear in this study, but an explanation could be due to hot and dry growing conditions during grain lling. Parameters of B-granules showed many signicant correlations with wheat and our properties, partly due to the inverse correlation between protein content and parameters of B-granules. There appears to be different optimum ranges of Bgranule weight % for ours with different protein contents. It seems that starch granule size distribution is a unique property that
HRW
Area Area Area Area Area Area Area Area
A AB A/our protein A B/our protein A AB A/our protein A B/our protein
** ** * *
*** *** *** * *** ***
HRS
105
affects physicochemical properties of wheat, our, and breadmaking properties in conjunction with its counterpart, protein. References
AACC, 2000. Approved Methods of the American Association of Cereal Chemists, 10th ed. AACC, St. Paul, MN, USA. Altenbach, S.B., DuPont, F.M., Kothari, K.M., Chan, R., Johnson, E.L., Lieu, D., 2003. Temperature, water and fertilizer inuence the timing of key events during grain development in a U.S. spring wheat. Journal of Cereal Science 37, 920. Bechtel, D.B., Wilson, J.D., 2000. Variability in a starch isolation method and automated digital image analysis system used for the study of starch size distributions in wheat our. Cereal Chemistry 77, 401405. Bechtel, D.B., Zayas, I., Kaleikau, L., Pomeranz, Y., 1990. Size-distribution of wheat starch granules during endosperm development. Cereal Chemistry 67, 5963. Bhullar, S.S., Jenner, C.F., 1985. Differential responses to high temperatures of starch and nitrogen accumulation in the grain of four cultivars of wheat. Australian Journal of Plant Physiology 12, 363375. Briarty, L.G., Hughes, C.E., Evers, A.D., 1979. The developing endosperm of wheat a stereological analysis. Annals of Botany 44, 641658. Brocklehurst, P.A., Evers, A.D., 1977. The size distribution of starch granules in endosperm of different sized kernels of the wheat cultivar Maris Huntsman. Journal of the Science of Food and Agriculture 28, 10841089. Chiotelli, E., Le Meste, M., 2002. Effect of small and large wheat starch granules on thermomechanical behavior of starch. Cereal Chemistry 79, 286293. Danno, G., Hoseney, R.C., 1982. Effects of dough mixing and rheologically active compounds on relative viscosity of wheat proteins. Cereal Chemistry 59,196198. DAppolonia, B.L., Gilles, K.A., 1971. Effect of various starches in baking. Cereal Chemistry 48, 625636. Dengate, H., Meredith, P., 1984. Variation in size distribution of starch granules from wheat grain. Journal of Cereal Science 3, 8390. Dengate, H.N., Baruch, D.W., McKenzie, J., Simmons, L.D., Meredith, P.W., Morrison, R., 1979. Fractionation of wheat starch on the basis of hydrated density. Starch 31, 4044. Dronzek, B.L., Hwang, P., Bushuk, W., 1972. Scanning electron microscopy of starch from sprouted wheat. Cereal Chemistry 49, 232239. Eliasson, A.C., 1989. Characterization of wheat starch and gluten as related to enduse properties. In: Pomeranz, Y. (Ed.), Wheat is Unique. AACC International, St. Paul, MN, pp. 177192. Evers, A.D., 1973. The size distribution among starch granules in wheat endosperm. Starch 25, 303304. Evers, A.D., Lindley, J., 1977. The particle-size distribution in wheat endosperm starch. Journal of Science and Food Agriculture 28, 98102. Gan, Z., Ellis, P.R., Schoeld, J.D., 1995. Mini review. Gas cell stabilization and gas retention in wheat bread dough. Journal of Cereal Science 21, 215230. Gibson, L.R., McCluskey, P.J., Tilley, K.A., Paulsen, G.M., 1998. Quality of hard red winter wheat grown under high temperature conditions during maturation and ripening. Cereal Chemistry 75, 421427. Gibson, L.R., Paulsen, G.M., 1999. Yield components of wheat grown under high temperature stress during reproductive growth. Crop Science 39, 18411846. Haraszi, R., Gras, P.W., Tomoskozi, S., Salgo, A., Bekes, F., 2004. Application of a micro Z-arm mixer to characterize mixing properties and water absorption of wheat our. Cereal Chemistry 81, 555560. Hayman, D., Sipes, K., Hoseney, R.C., Faubion, J.M., 1998. Factors controlling gas cell failure in bread dough. Cereal Chemistry 75, 585589. Hoseney, R.C., Finney, K.F., Pomeranz, Y., Shogren, M.D., 1971. Functional (breadmaking) and biochemical properties of wheat our components. VII. Starch. Cereal Chemistry 48, 191201. Hughes, C.E., Briarty, L.G., 1976. Stereological analysis of the contribution to mature wheat endosperm starch by large and small granules. Starch 28, 336337. Karlsson, R., Olered, R., Eliasson, A.-C., 1983. Changes in starch granule size distribution and starch gelatinization properties during development and maturation of wheat, barley and rye. Starch 35, 335340. Kulp, K., 1973. Characteristics of small-granule starch of our and wheat. Cereal Chemistry 50, 666679. Lelievre, J., Lorenz, K., Meredith, P., Baruch, D.W., 1987. Effects of starch particle size and protein concentration on breadmaking performance. Starch 39, 347352. Lineback, D.R., 1984. The starch granule organization and properties. Bakers Digest 58 (2), 16. 1821. MacLeod, L.C., Duffus, C.M., 1988. Temperature effects on starch granules in developing barley grain. Journal of Cereal Science 8, 2937. MacRitchie, F., 1992. Physicochemical properties of wheat proteins in relation to functionality. Advances in Food Research 36, 185. Maghirang, E.B., Lookhart, G.L., Bean, S.R., Pierce, R.O., Xie, F., Caley, M.S., Wilson, J.D., Seabourn, B.W., Ram, M.S., Park, S.H., Chung, O.K., Dowell, F.E., 2006. Comparison of quality characteristics and breadmaking functionality of hard red winter and hard red spring wheat. Cereal Chemistry 83, 520528. Meredith, P., 1981. Large and small starch granules in wheat are they really different? Starch 33, 4044. Miller, K.A., Hoseney, R.C., 1999. Dynamic rheological properties of wheat starch gluten doughs. Cereal Chemistry 76, 105109. Morrison, W.R., Gadan, H., 1987. The amylose and lipid contents of starch granule in developing wheat endosperm. Journal of Cereal Science 5, 263275.
Morrison, W.R., Scott, D.C., 1986. Measurements of the dimensions of wheat starch granule populations using a Coulter Counter with 100-channel analyzer. Journal of Cereal Science 4, 1321. Nikuni, Z., 1978. Studies on starch granules. Starch 30, 105111. Ohm, J.B., Chung, O.K., 1999. Gluten, pasting, and mixograph parameters of hard winter wheat ours in relation to breadmaking. Cereal Chemistry 76, 606613. Ohm, J.B., Chung, O.K., Deyoe, C.W., 1998. Single kernel characteristics of hard winter wheats in relation to milling and baking quality. Cereal Chemistry 75, 156161. Park, S.H., Chung, O.K., Seib, P.A., 2004. Size distribution and properties of wheat starch granules in relation to crumb grain score of pup-loaf bread. Cereal Chemistry 81, 699704. Park, S.H., Chung, O.K., Seib, P.A., 2005. Effects of varying weight ratios of large and small wheat starch granules on experimental straight-dough bread. Cereal Chemistry 82, 166172. Park, S.H., Chung, O.K., Seib, P.A., 2006. Levels of protein and protein composition in hard winter wheat ours and the relationship to breadmaking. Cereal Chemistry 83, 418423. Parker, M.L., 1985. The relationship between A-type and B-type starch granules in the developing endosperm of wheat. Journal of Cereal Science 3, 271278. Peng, M., Gao, M., Abdel-Aal, E.-S.M., Hucl, P., Chibbar, R.N., 1999. Separation and characterization of A- and B-type starch granules in wheat endosperm. Cereal Chemistry 76, 375379. Petrofsky, K.E., Hoseney, R.C., 1995. Rheological properties of dough made with starch and gluten from several sources. Cereal Chemistry 72, 5358. Pomeranz, Y., 1988. Chemical composition of kernel structures. In: Pomeranz, Y. (Ed.), Wheat Chemistry and Technology, vol. 1. AACC International, St. Paul, MN, pp. 97158. Raeker, M.O., Gaines, C.S., Finney, P.L., Donelson, T., 1998. Granule size distribution and chemical composition of starches from 12 soft wheat cultivars. Cereal Chemistry 75, 721728. Sandstedt, R.M., 1961. The function of starch in the baking of bread. Bakers Digest 35 (3), 3642. 44. Sahlstrom, S., Brathen, E., Lea, P., Autio, K., 1998. Inuence of starch granule size distribution on bread characteristics. Journal of Cereal Science 28, 157164. Seib, P.A., 1994. Wheat starch: isolation, structure, and properties. Oyo Toshitsu Kagaku 41, 4969. Shi, Y.C., Seib, P.A., Bernardin, J.E., 1994. Effects of temperature during grain-lling on starches from six wheat cultivars. Cereal Chemistry 71, 369383. Skerritt, J.H., Hac, L., Lindsay, M.P., Bekes, F., 1999. Depolymerization of the gluten macropolymer during mixing: II. Differences in retention of specic glutenin subunits. Cereal Chemistry 76, 402409. Simmonds, D.H., OBrien, T.P., 1981. Morphological and biochemical development of wheat endosperm. Advances in Cereal Science and Technology 4, 514. Soh, H.N., Sissons, M.J., Turner, M.A., 2006. Effects of starch granule size distribution and elevated amylose content on durum dough rheology and spaghetti cooking quality. Cereal Chemistry 83, 513519. Sollars, W.F., Rubenthaler, G.L., 1971. Performance of wheat and other starches in reconstituted ours. Cereal Chemistry 48, 397410. Soulaka, A.B., Morrison, W.R., 1985a. The amylose and lipid contents, dimensions, and gelatinization characteristics of some wheat starches and their A- and Bgranule fractions. Journal of the Science of Food and Agriculture 36, 709718. Soulaka, A.B., Morrison, W.R., 1985b. The bread baking quality of six wheat starches differing in composition and physical properties. Journal of the Science of Food and Agriculture 36, 719727. Stoddard, F.L., 1999. Survey of starch particle-size distribution in wheat and related species. Cereal Chemistry 76, 145149. Stone, P.J., Nicolas, M.E., 1994. Wheat cultivars vary widely in their responses of grain yield and quality to short periods of post-anthesis heat stress. Australian Journal of Plant Physiology 21, 887900. Tanaka, K., Bushuk, W., 1973. Changes in our proteins during dough mixing. II. Gel ltration and electrophoresis results. Cereal Chemistry 50, 597605. Tashiro, T., Wardlaw, I.F., 1990. The effects of high temperature at different stages of ripening on grain set, grain weight, and grain dimension in the semi-dwarf wheat Banks. Annals of Botany 65, 5161. Tester, R.F., South, J.B., Morrison, W.R., Ellis, R.P., 1991. The effects of ambient temperature during the grain-lling period on the composition and properties of starch from four barley genotypes. Journal of Cereal Science 13, 113127. U.S. Wheat Associate, 2002. Harvest Report October 8, 2002 nal. http://www. uswheat.org/harvestReports/doc/3DCEBA32BAAE4ACF85256F8100572BE1?Open Document# (accessed April 24, 2008). Van Vliet, T., Janssen, A.M., Bloksma, A.H., Walstra, P., 1992. Strain hardening of dough as a requirement for gas retention. Journal of Texture Studies 23, 439460. Weegels, P.L., van de Pijpekamp, A.M., Graveland, A., Hamer, R.J., Schoeld, J.D., 1996. Depolymerisation and repolymerisation of wheat glutenin during dough processing. I. Relationships between glutenin macropolymer content and quality parameters. Journal of Cereal Science 23, 103111. Wilson, J.D., Bechtel, T.C., Seib, P.A., 2006. Measurement of wheat starch granule size distribution using image analysis and laser diffraction technology. Cereal Chemistry 83, 259268. Wong, R.B.K., Lelievre, J., 1982. Comparison of the crystallinities of wheat starches with different swelling capacities. Starch 34, 159161.

Total phenolics, avonoids, antioxidant capacity in rice grain and their relations to grain color, size and weight
Yun Shen, Liang Jin, Peng Xiao, Yan Lu, Jinsong Bao*
Institute of Nuclear Agricultural Sciences, Key Laboratory of Chinese Ministry of Agriculture and Zhejiang Province for Nuclear-Agricultural Sciences, College of Agriculture and Biotechnology, Zhejiang University, Hua Jiachi Campus, Hangzhou 310029, Peoples Republic of China
Article history: Received 2 March 2008 Received in revised form 14 June 2008 Accepted 7 July 2008 Keywords: Antioxidant capacity Flavonoid Phenolics Rice
a b s t r a c t
Total phenolics, avonoid contents and antioxidant capacity from a wide collection of rice germplasm were measured, and their relations to grain color, grain size and 100-grain weight were investigated. Highly signicant genotypic differences were observed in total phenolics, avonoid contents and 2,2azino-bis-(3-ehylbenzothiazoline-6-sulphonic acid) diammonium salt (ABTS) radical cation antioxidant capacity. They displayed an increasing order in the white rice, red rice and black rice, yet several white rice had higher phenolics and avonoids contents than the red rice. Signicant positive pair-wise correlations were found among the phenolics, avonoid contents and antioxidant capacity, and the coefcient between the phenolic contents and antioxidant capacity was extremely high (r 0.96). Among all rice accessions, the grain color parameters had negative correlations with the phenolics, avonoid contents and antioxidant capacity (p < 0.001). The negative correlation between a* and antioxidant capacity, and the positive correlation between H and antioxidant capacity were consistent within the respective white rice and red rice groups. Flavonoid contents had positive correlation with grain length and length to width ratio, and had negative correlation with the 100-grain weight among all rice accessions. It was also found that 100-grain weight still had negative correlations with phenolics, avonoid contents and antioxidant capacity within the white rice genotypes. These relationships may serve as indexes to indirectly select breeding lines high in the phenolics, avonoids and antioxidant capacity. Principal component analysis including the information for phenolics, avonoids, antioxidant capacity, grain color parameters, grain size and 100-grain weight extracted ve principal components that explained 83.7% of the total variances. The results of this study may provide new opportunities for rice breeders and eventually commercial rice growers to promote the production of rice with enhanced nutritional quality. 2008 Elsevier Ltd. All rights reserved.
1. Introduction Rice is a staple food being consumed by nearly half of the world population. Nutritional quality of rice has received more attention in the developing countries, where monotonous consumption of rice may lead to deciencies of essential minerals, vitamins, and other nutritional compositions (Bouis et al., 2003). This is not caused by nutritional deciency in rice grain itself, but due to it being traditionally eaten in the form of the milled white kernel. Milling of the brown rice to obtain milled rice removes bran layers
Abbreviations: ABTS, 2,2-azino-bis-(3-ehylbenzothiazoline-6-sulphonic acid) diammonium salt; GAE, gallic acid equivalent; RE, rutin equivalent; TEAC, trolox equivalent antioxidant capacity. * Corresponding author. Tel.: 86 571 8697 1932; fax: 86 571 8697 1421. E-mail address: jsbao@zju.edu.cn (J. Bao). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.07.010
that are rich in protein, ber, oil, minerals, vitamins, and other phytochemicals (Orthoefer and Eastman, 2004; Yokoyama, 2004), leading to loss of most of the nutritional components of the rice grain. Numerous studies have shown that the essential phytochemicals in fruits, vegetables and cereal grains, including rice, are signicantly associated with reduced risk of developing chronic diseases such as cardiovascular disease, type 2 diabetes, and some cancers (Liu, 2007; Yawadio et al., 2007; Yokoyama, 2004). Biofortication of rice to improve nutritional quality and to combat nutritional deciency by a transgenic engineering approach has several successful examples (Bouis et al., 2003). Ye et al. (2000) introduced the b-carotene synthesis pathway to rice endosperm by genetic engineering to obtain the golden rice that produced 1.6 mg/g of b-carotene in the grain. Storozhenko et al. (2007) reported biofortication of folate content in rice grain by over expression of two Arabidopsis genes encoding GTP cyclohydrolase I and aminodeoxychorismate synthase under the
Y. Shen et al. / Journal of Cereal Science 49 (2009) 106111
107
control of strong endosperm-specic promoters. The transgenic rice grain contained up to 100-fold higher folate levels compared to the wild type. It should be noted that conventional breeding is still possible to improve the nutritional components in grains. For example, Harjes et al. (2008) reported that variations at the lycopene epsilon cyclase (lcyE) locus alter ux down a-carotene versus b-carotene branches of the carotenoid pathways in the maize grain, and four lcyE natural polymorphism explained 58% of the variation in these two branches and a threefold difference in provitamin A compounds. Thus, selection of favorable lcyE alleles with inexpensive molecular markers enables developing-country breeders to more effectively improve provitamin A levels. However, effort towards biofortication of rice grain to improve nutritional quality by conventional breeding has been scarcely reported (Bouis et al., 2003). As a primary step to achieve this goal, it is necessary to investigate the genotypic diversity in the phytochemicals among diverse rice accessions, so as to nd a way to enrich these compositions by breeding. The genotypic diversity of some phytochemicals in rice bran layers has been widely characterized (Dykes and Rooney, 2007; Liu, 2007). For example, Bergman and Xu (2003) reported genotypic and environmental effects on tocopherol, tocotrienol, and g-oryzanol contents of rice, and Miller and Engel (2006) also reported the contents of g-oryzanol in brown rice. Jiang et al. (2007) reported minerals contents and their correlation with other quality traits of rice. However, some phytochemicals, including phenolics and avonoids, have not received as much attention as other compositions in rice grains and the phytochemicals in other cereals, fruits and vegetables (Liu, 2007). Phenolics are compounds possessing one or more aromatic rings with one or more hydroxyl groups (Liu, 2007). Phenolic compounds in diet may provide health benets associated with reduced risk of chronic disease (Liu, 2007). Chinese medicinal plants have high levels of phenolics and potent antioxidant capacity, which might contribute to the protective effects against cancer (Cai et al., 2004). In rice, Goffman and Bergman (2004) studied the genotypic and environmental effects of the kernel phenolic content, and found that bran color was highly statistically signicant for bran phenolic contents. Flavonoids are one group of phenolics, which consists of two aromatic rings linked by 3 carbons that are usually in an oxygenated heterocycle ring (Liu, 2004). Anthocyanins are a group of reddish to purple watersoluble avonoids that are the primary pigments in the red and black grains, and have been widely identied and characterized in cereal grains (Abdel-Aal et al., 2006). The major components of anthocyanidins in colored rice are cyaniding-3-O-b-glucoside and peonidin-3-O-b-glucoside (Abdel-Aal et al., 2006; Yawadio et al., 2007). There have been few reports on characterization of other avonoids such as avonols, avones, avanols, and avanones. The phenolic compounds are also known as antioxidants (Abdel-Aal et al., 2006; Adom and Liu, 2002; Hu et al., 2003). Antioxidants have long been recognized to have protective functions against oxidative damage, and are associated with reduced risk of chronic diseases (Adom and Liu, 2002; Liu, 2007). Other phytochemicals such as carotenoids, tocols and g-oryzanols are also antioxidants (Aguilar-Garcia et al., 2007; Choi et al., 2007; Xu et al., 2001). The objective of this study was to evaluate total phenolics, avonoids and antioxidant capacity of a large number of rice genotypes (481 accessions) and to analyze their relationships with grain color, size and 100-grain weight. The results of this study could provide rice breeders and eventually commercial rice growers new opportunities to promote the production of rice with enhanced levels of the bioactive compounds.
2. Experimental 2.1. Materials A total of 481 rice accessions including 423 white rice, 52 red rice and 6 black rice, were employed in this study. All the rice was grown in the Zhejiang University farm in 2006. They were sown in late May, transplanted on June 20, and harvested in October, and the eld management followed conventional practices. Rice grains were air-dried and stored at room temperature for three months. Then they were dehusked on a Satake Rice Machine (Satake Co., Japan), and ground to pass through a 100-mesh sieve on a Cyclone Sample Mill (UDY Corporation, Fort Collins, Colorado, USA). 2,2Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS), potassium persulfate and gallic acid were purchased from BBI (Ontario, Canada), 6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid (Trolox) from Sigma/Aldrich (St. Louis, MO), FolinCiocalteu reagent from Fluka Chemie AG (Buchs, Switzerland), and rutin from Oukang Company (Chengdu, China). 2.2. Extraction Wholemeal ours (1 g) of each accession were extracted with 25 mL of methanol containing 1% HCl for 24 h at 24 C. The procedure was repeated twice. The methanolic extracts were centrifuged at w4000g for 15 min and the supernatants were pooled and stored at 4 C. 2.3. Color of rice grain The color of rice grain sample was measured with a TC-PIIG automatic color difference meter (Beijing Optical Instrument Factory, Beijing, China). Color measurements were expressed as tristimulus parameters, L*, a*, and b*. L* indicates lightness (100 white and 0 black). a* indicates rednessgreenness and b* indicates yellownessblueness. In addition, the chroma (C) value indicates color intensity or saturation, calculated as C a*2 b*2 1=2 , and Hue angle was calculated as H tan1 (b*/ a*) (Bao et al., 2005). 2.4. Total phenolics Total phenolic content was assayed by the FolinCiocalteu colorimetric method with slight modication (Bao et al., 2005; Cai
Table 1 Variations in phenolics, avonoids contents and antioxidant capacity among white (n 423), red (n 52) and black (n 6) rice genotypes Phenolicsa Total rice Mean SD CV (%) Range White rice Mean SD CV (%) Range Red rice Mean SD CV (%) Range Black rice Mean SD CV (%) Range 197.5 144.8 73.3 108.11244.9 151.8 19.5 12.9 108.1251.4 470.1 107.2 22.8 165.8731.8 1055.7 176.2 16.7 841.01244.9 Flavonoidsa 134.7 19.8 14.7 88.6286.3 131.6 14.2 10.8 88.6170.7 147.2 18.0 12.3 108.7190.3 240.6 38.1 15.8 187.6286.3 Antioxidant capacitya 0.413 0.696 168.63 0.0125.533 0.196 0.073 37.33 0.0120.413 1.705 0.600 35.22 0.2912.963 4.484 1.095 24.41 2.5275.533
a Phenolics content was expressed as mg GAE/100 g, avonoids content was expressed as mg RE/100 g, and antioxidant capacity was expressed as mM TAEC.
108
250 black 200 red white
2.5. Total avonoids Total avonoid content was determined by a colorimetric method (Bao et al., 2005) with minor modication. Aliquots (0.5 mL) of appropriately diluted extracts or standard solutions were pipetted into 15-mL polypropylene conical tubes containing 2 mL double distilled H2O and mixed with 0.15 mL 5% NaNO2. After 5 min, 0.15 mL 10% AlCl3$6H2O solution was added, and the mixture was allowed to stand for another 5 min, and then 1 mL 1 M NaOH was added. The reaction solution was well mixed, kept for 15 min, and the absorbance was determined at 415 nm. Total avonoid content was calculated using the standard rutin curve, and expressed as mg rutin equivalent (mg RE) per 100 g of dry weight. 2.6. Radical cation ABTS scavenging activity

No. of accessions
150
100
50
<150
150-300
300-450
450-600
>600
Phenolics content (mg GAE / 100g)

300 black 250 red white 200 150 100 50
Total antioxidant capacity of rice extracts was carried out using a spectrophotometer by the improved 2,2-azino-bis(3-ehylbenzothiazoline-6-sulphonic acid) diammonium salt (ABTS) radical cation method as described (Bao et al., 2005; Cai et al., 2004). ABTS solution (3.9 mL, absorbance of 0.700) was added to 0.1 mL of the extracts and mixed thoroughly. The reaction mixture was kept at room temperature for 6 min and the absorbance was immediately recorded at 734 nm. Trolox standard solution in 80% ethanol was prepared and assayed under the same conditions. Results were expressed in terms of Trolox equivalent antioxidant capacity (TEAC, mM Trolox equivalents per 100 g dry weight). 2.7. Statistical analysis
No. of accessions
<100
100-120
120-140
140-160
160-180
>180
Flavonoids content (mg RE / 100g)

450 400 350 black red white
All the analyses were carried out at least in duplicate and in randomized order with mean values being reported. Analysis of variance (ANOVA), correlation analysis and principal component analysis of the results were performed in SAS (Software Version 9.1. SAS Institute Inc., Cary, NC). 3. Results 3.1. Total phenolics, avonoid contents and antioxidant capacity There were wide range of variations in the total phenolics in rice grain (Table 1, Fig. 1). Among all the rice accessions, total phenolic content ranged from 108.1 to 1244.9 mg GAE/100 g, with the lower values coming from the white rice, while the higher values were from red and black rice (Fig. 1). Variations were still found within the white rice and red rice, ranging from 108 to 251 mg GAE/100 g and from 165.8 to 731.8 mg GAE/100 g for white and red rice, respectively (Table 1). However, several red rice accessions still had lower total phenolic contents than the white rice (Fig. 1). Flavonoid contents in all the rice ranged from 88.6 to 286.3 mg RE/100 g. The mean avonoid contents among the white, red and black rice were 131.6, 147.2 and 240.6 mg RE/100 g, respectively. Even though red rice had average higher levels of avonoids than the white rice, some red rice accessions still had
No. of accessions
300 250 200 150 100 50 0 <1 1.0-2.0 2.0-3.0 >3.0
Antioxidant capacity (mM TAEC)

Fig. 1. Mean distributions of phenolics, avonoids contents and antioxidant capacity among white rice (423), red rice (52) and black rice (6) accessions.
et al., 2004). Briey, aliquots (1.0 mL) of appropriately diluted extracts or standard solutions were mixed with 0.5 mL 0.5 N FolinCiocalteu reagent, then the reaction was neutralized with saturated sodium carbonate (75 g/L). The absorbance of the resulting blue color was recorded using a spectrophotometer after incubation for 2 h at 23 C. A calibration curve was prepared using gallic acid solution. Total phenolics contents were expressed as milligrams of gallic acid equivalent (mg GAE) per 100 g of dry weight.
Table 2 Pair-wise correlations among phenolics, avonoids contents and antioxidant capacity (ABTS) among the total rice, white and red rice genotypes Total rice Flavonoids Phenolics Flavonoids 0.681*** ABTS 0.962*** 0.612*** White rice Flavonoids 0.703*** ABTS 0.231*** 0.101* Red rice Flavonoids 0.461*** ABTS 0.777*** 0.342*
*, ** and *** were signicant at 0.05, 0.01 and 0.001 probability level, respectively.
109
70 60 50 40 30 20 10 0 -10 -20 -30

Fig. 2. Mean color parameters of white, red and black rice.
L* a* b*
white rice
red rice
black rice
lower contents than white rice, whereas most of red rice had lower contents than the black rice (Fig. 1). The total antioxidant capacity was measured using the ABTS assay. It was varied to a great extent, averaged 0.413 mM TEAC, ranging from 0.012 to 5.533 mM TEAC among the total rice accessions (Table 1, Fig. 1). Among the white rice, the mean ABTS was 0.196 mM TEAC, ranging from 0.012 to 0.413 mM TEAC, whereas among the red rice, it averaged 1.705 mM TEAC, ranging from 0.291 to 2.963 mM TEAC. The six black rice samples had average antioxidant capacity of 4.484 mM TEAC, around three times of that of the red rice (Table 1). 3.2. Correlations among total phenolics, avonoids and antioxidant capacity Pair-wise correlations between the three parameters were positive (P < 0.001) among all rice groups, which were still true within the respective white rice and red rice accessions (Table 2). The phenolic contents were highly positively correlated with the antioxidant capacity (r 0.962) among all the rice. However, the correlation coefcient within the white rice accessions (r 0.231) was much smaller than that within the red rice accessions (r 0.777). The correlation coefcient between phenolic contents and avonoid contents was higher within white rice groups (r 0.703) than that within the red rice accessions (r 0.461). The correlation coefcient between avonoid content and antioxidant capacity was much smaller within the white and red rice accessions (Table 2) than that within the total rice groups (r 0.612), which might be due to the smaller variation in each group when compared to the total rice accessions (Table 2). 3.3. Relationships between phenolics, avonoids, antioxidant capacity and grain color Apparently, the red rice had L*, a*, and b* values of color parameters smaller than the white rice, but larger than the black rice, though the color parameters still differed within the white and
red rice groups (Fig. 2). Among total rice accessions, the ve color parameters (L*, a*, b*, C and H ) were signicantly negatively correlated with phenolics, avonoid contents, and antioxidant capacity (Table 3). Within the white rice accessions, some correlations were still signicant (Table 3), but only the L* vs phenolic contents, a* vs antioxidant capacity, H vs phenolic contents and avonoid contents were negatively correlated, whereas antioxidant capacity vs L*, b* and H , and a* vs phenolics and avonoid contents were positive, which implied that the correlation from total rice accessions could not be applied to the individual rice groups. Within the red rice group, a* was still negatively correlated with phenolic contents and antioxidant capacity, C was also negatively correlated with antioxidant capacity, but the H was positively correlated with phenolic contents and antioxidant capacity (Table 3). It was worthy to note that the negative correlation between a* and antioxidant capacity was consistent in white and red rice groups, and so was the positive correlation between H and antioxidant capacity (Table 3), suggesting that these correlations could be used as indirect indexes to select rice breeding lines with high antioxidant capacity. 3.4. Relationships between phenolics, avonoids, antioxidant capacity and grain size and 100-grain weight Among all rice accessions, only avonoid contents were poorly positively correlated with grain length, length/width ratio and 100grain weight (Table 4). However, it seems that the antioxidant capacity was negatively correlated with, whereas the avonoids contents were positively correlated with grain length and length/ width ratio within the white rice group (Table 4). The 100-grain weight was negatively correlated with phenolics, avonoid contents and antioxidant capacity. Interestingly, the avonoid contents still had positive correlation with grain length within the red rice group (r 0.325). Contrast to the relationship within the total rice and white rice, the 100-grain weight was positively correlated with the avonoid contents within the red rice accessions (Table 4). 3.5. Principal component analysis Principal component analysis was performed on the ten variables including phenolic contents, avonoid content, antioxidant capacity, color parameters, grain size and 100-grain weight (Tables 5 and 6). All the twelve principal components and their corresponding eigenvalues and variances are listed in Table 5. The results indicated that the rst ve principal components could explain 83.7% of total variance. The rst principal component (PC1) was the most important one, explaining 26.1% of the total variance (Table 5). The PC1 represented the phenolics content, C and H of color parameters (Table 6). The second principal component (PC2) accounted for an additional 19.2% of the total variances, which was mainly attributed to avonoid contents, L* and a* of color parameters. The third principal component (PC3) accounted for 16.2% of the total variance; the variation was mainly contributed by
Table 3 Correlation coefcients between phenolics, avonoids contents, antioxidant capacity (ABTS) and grain color parameters among the total rice, white and red rice genotypes Total rice Phenolics L* a* b* C H 0.774*** 0.349*** 0.592*** 0.271*** 0.589*** Flavonoids 0.463*** 0.277*** 0.481*** 0.145** 0.529*** ABTS 0.738*** 0.398*** 0.579*** 0.311*** 0.582*** White rice Phenolics 0.103* 0.170*** 0.066 0.156** 0.152** Flavonoids 0.065 0.136** 0.018 0.089 0.126* ABTS 0.333*** 0.460*** 0.212*** 0.001 0.477*** Red rice Phenolics 0.166 0.292* 0.161 0.191 0.326* Flavonoids 0.014 0.165 0.101 0.089 0.195 ABTS 0.128 0.335* 0.261 0.307* 0.332*
110
Table 4 Correlation coefcients between phenolics, avonoids contents, antioxidant capacity (ABTS) and grain size and 100-grain weight among the total rice, white and red rice genotypes Total rice Phenolics Length Width Length/Width ratio 100 grain weight 0.005 0.011 0.031 0.067 Flavonoids 0.169*** 0.030 0.106* 0.110* ABTS 0.036 0.007 0.042 0.065 White rice Phenolics 0.064 0.042 0.014 0.146** Flavonoids 0.177*** 0.052 0.132** 0.117* ABTS 0.178*** 0.054 0.097* 0.160** Red rice Phenolics 0.072 0.021 0.064 0.126 Flavonoids 0.325* 0.155 0.083 0.322* ABTS 0.106 0.068 0.113 0.168
antioxidant capacity, grain length and grain width. The fourth principal component mainly representing the grain length to width ratio and 100-grain weight explained 13.2% of the total variance. The fth principal component representing the b* of color parameter with eigenvalue of 1.095, explained an additional 9.1% of the total variance (Tables 5 and 6).
4. Discussion The role of phenolics as natural antioxidants has attracted considerable interest due to their pharmacological functions. Increased consumption of phenolic compounds has been associated with the reduced risk of cardiovascular diseases and certain cancers (Liu, 2004, 2007; Dykes and Rooney, 2007). The whole rice grain had phenolic contents ranging from 108.1 to 1244.9 mg GAE/ 100 g (Table 1), depending on the color of grain (Choi et al., 2007; Goffman and Bergman, 2004). Goffman and Bergman (2004) reported that the phenolic contents in the white, red and purple rice ranged from 25 to 246, 34 to 424, 69 to 535 mg GAE/100 g, which was a little lower than this study (Table 1). Three reasons may explain the differences. First, they used rice materials grown in two years, which means that material harvested from the rst year had to be stored. Storage of rice grain results in a decrease of phenolic content (Zhou et al., 2004), thus their data was lower than ours. Second, rice phenolic compounds exist in free, esteried and insoluble-bound forms, and insoluble-bound phenolics may be released by base, acid or enzymatic treatment of samples prior to extraction (Adom and Liu, 2002; Choi et al., 2007; Zhou et al., 2004). Our extraction solution including 1% HCL may cause release of at least part of the bound phenolics, thus leading to higher phenolic contents. Third, more rice accessions were used in this study may provide more wide diversity in the phenolic contents. Flavonoids have potent antioxidant and anticancer activities (Adom and Liu, 2002; Dykes and Rooney, 2007; Hu et al., 2003). Coloration of rice may be derived from accumulation of anthocyanins (Furukawa et al., 2007). Thus, it could be expected that the white rice had mean avonoid content (131.6 mg RE/100 g) lower than those of red rice (147.2 mg RE/100 g) and black rice (240 mg RE/100 g)
(Table 1). It should be noted that some red rice accessions were lower in the avonoid contents than white rice (Fig. 1). It was possible that other avonoid compositions (e.g. avonols, avanones) rather than anthocyanins in the white rice were higher than those in the red rice. The antioxidant capacity of rice grain was contributed not only by phenolic compounds, but also by other phytochemicals, such as carotenoids, tocols and g-oryzanols (Choi et al., 2007; Xu et al., 2001). However, phenolic contents were reportedly positively correlated with TABS antioxidant capacity (r 0.962), which was consistent with nearly all other studies (Adom and Liu, 2002; Choi et al., 2007) using samples from different plant origins (Cai et al., 2004). Biofortication by conventional breeding is one of the strategies to improve nutritional quality of rice grain. It is apparent that phenolics, avonoids and antioxidant capacity differ from the grain color (Table 1), yet these components and antioxidant capacity still differ among the white color rice (Table 1). Therefore, breeding efforts for better nutritional quality may be applied for white rice, if the whole grain is used for food product development. For ease of breeding, indirect selection approaches are often used in the breeding program. One of indirect selection methods is to use the color parameters which can be easily measured in an automatic color difference meter. The color parameters L*, b* and H are positively, while a* is negatively correlated with antioxidant capacity (Table 3). Thus, breeding lines high in antioxidant could be indirectly selected by selecting grain with larger L*, b* and H , but smaller a*. It is well known that most of the phytochemicals are rich in the bran layers which make up approximately 810% of the rough rice weight (Yokoyama, 2004), though some of them are still present in the milled rice. It is general knowledge that, in a given weight of rice, the larger the grain size, the smaller surface area per weight of grain. Smaller surface area might imply lower phytochemicals in the bran layers. However, whether the grain size negatively correlated with phenolics, avonoids and antioxidant capacity was not tested before. It is found that the 100-grain weight was truly negatively correlated with avonoid contents among the total rice accessions, and the correlations were still true among white rice, but not true among red rice accessions (Table 4). The grain size was
Table 5 Principal component analysis for all the ten parameters Component 1 2 3 4 5 6 7 8 9 10 11 12 Eigenvalue 3.126 2.300 1.941 1.579 1.095 0.726 0.691 0.263 0.215 0.037 0.023 0.004 Variance (%) 26.05 19.17 16.18 13.16 9.12 6.05 5.76 2.19 1.79 0.30 0.19 0.03 Cumulative variance (%) 26.05 45.22 61.40 74.56 83.68 89.73 95.49 97.68 99.47 99.78 99.97 100.00
Table 6 Sources of variation for the rst ve principal components (PC) PC1 Phenolics Flavonoids ABTS L* a* b* C H Length Width Length/width ratio 100 grain weight 0.481 0.143 0.323 0.224 0.173 0.274 0.485 0.464 0.076 0.128 0.106 0.073 PC2 0.065 0.591 0.237 0.479 0.548 0.098 0.011 0.062 0.079 0.128 0.101 0.123 PC3 0.092 0.187 0.433 0.320 0.255 0.087 0.297 0.313 0.492 0.394 0.095 0.049 PC4 0.075 0.136 0.029 0.094 0.135 0.018 0.111 0.095 0.305 0.364 0.522 0.655 PC5 0.215 0.164 0.333 0.240 0.210 0.564 0.109 0.062 0.455 0.413 0.030 0.053
111
not necessarily negatively correlated with avonoids, but inversely the relationship was positive. The antioxidant capacity was negatively correlated with grain length, length/width ratio and 100grain weight (Table 4), so in the white rice breeding practice, high antioxidant capacity of breeding lines could be indirectly selected with short grain length and smaller 100-grain weight in breeding programs. In conclusion, this study found wide diversity in the phenolics, avonoid contents and antioxidant capacity in the whole rice grain. These data provide opportunities to further increase the content of phenolics, avonoids and antioxidant capacity by breeding, especially in white rice. Their relationships with grain color, grain size and 100-grain weight could serve as indexes to indirectly select rice breeding lines high in phenolics, avonoids and antioxidant capacity. The results could provide rice breeders and eventually commercial rice producers, with new opportunities to promote the production of rice with enhanced levels of the phytochemicals. Rice genotypes rich in phytochemicals may be incorporated into functional foods. Acknowledgement The authors thank Mr. Junquan Yu and Dr. Jianliang Lu for their assistance in chemical analysis and the color parameters test, respectively. We also thank the anonymous reviewers for their constructive comments. Financial support for this work was provided in part by National High Technology Development Project, National Natural Science Foundation of China and the Science and Technology Department of Zhejiang Province. References
Abdel-Aal, E.M., Young, J.C., Rabalski, I., 2006. Anthocyanin composition in black, blue, pink, purple, and red cereal grains. Journal of Agricultural and Food Chemistry 54, 46964704. Adom, K.K., Liu, R.H., 2002. Antioxidant activity of grains. Journal of Agricultural and Food Chemistry 50, 61826187. Aguilar-Garcia, C., Gavino, G., Baragano-Mosqueda, M., Hevia, P., Gavino, V.C., 2007. Correlation of tocopherol, tocotrienol, g-oryzanol and total polyphenol content in rice bran with different antioxidant capacity assays. Food Chemistry 102, 12281232. Bao, J.S., Cai, Y., Sun, M., Wang, G.Y., Corke, H., 2005. Anthocyanins, avonols, and free radical scavenging activity of Chinese bayberry (Myrica rubra) extracts and their color properties and stability. Journal of Agricultural and Food Chemistry 53, 23272332.
Bergman, C.J., Xu, Z., 2003. Genotype and environment effects on tocopherol, tocotrienol, and g-oryzanol contents of Southern U.S. rice. Cereal Chemistry 80, 446449. Bouis, H.E., Chassy, B.M., Ochanda, J.O., 2003. Genetically modied food crops and their contribution to human nutrition and food quality. Trends in Food Science and Technology 14, 191209. Cai, Y.Z., Luo, Q., Sun, M., Corke, H., 2004. Antioxidant activity and phenolic compounds of 112 traditional Chinese medicinal plants associated with anticancer. Life Science 74, 21572184. Choi, Y., Jeong, H.S., Lee, J., 2007. Antioxidant activity of methanolic extracts from some grains consumed in Korea. Food Chemistry 103, 130138. Dykes, L., Rooney, L.W., 2007. Phenolic compounds in cereal grains and their health benets. Cereal Foods World 52, 105111. Furukawa, T., Maekawa, M., Oki, T., Suda, I., Lida, S., Shimada, H., Takamure, I., Kadowaki, K.I., 2007. The Rc and Rd genes are involved in proanthocyanidin synthesis in rice pericarp. Plant Journal 49, 91102. Goffman, F.D., Bergman, C.J., 2004. Rice kernel phenolic content and its relationship with antiradical efciency. Journal of the Science of Food and Agriculture 84, 12351240. Harjes, C.E., Rocheford, T.R., Bai, L., Brutnell, T.P., Kandianis, C.B., Sowinski, S.G., et al., 2008. Natural genetic variation in Lycopene Epsilon Cyclase tapped for maize biofortication. Science 319, 330333. Hu, C., Zawistowski, J., Ling, W., Kitts, D.D., 2003. Black rice (Oryza sativa L. indica) pigmented fraction suppresses both reactive oxygen species and nitric oxide in chemical and biological model systems. Journal of Agricultural and Food Chemistry 51, 52715277. Jiang, S.L., Wu, J.G., Feng, Y., Yang, X.E., Shi, C.H., 2007. Correlation analysis of mineral element contents and quality traits in milled rice (Oryza sativa L.). Journal of Agricultural and Food Chemistry 55, 96089613. Liu, R.H., 2004. Potential synergy of phytochemicals in cancer prevention: mechanism of action. Journal of Nutrition 134, 3479S3485S. Liu, R.H., 2007. Whole grain phytochemicals and health. Journal of Cereal Science 46, 207219. Miller, A., Engel, K.H., 2006. Content of g-oryzanol and composition of steryl ferulates in brown rice (Oryza sativa L.) of European origin. Journal of Agricultural and Food Chemistry 54, 81278133. Orthoefer, F.T., Eastman, J., 2004. Rice bran and oil. In: Champagne, E.T. (Ed.), Rice Chemistry and Technology. AACC, St. Paul, Minnesota, U.S.A., pp. 569593. Storozhenko, S., Brouwer, V.E., Volckaert, M., Navarrete, O., Blancquaert, D., Zhang, G.F., et al., 2007. Folate fortication of rice by metabolic engineering. Nature Biotechnology 25, 12771279. Xu, Z., Hua, N., Godber, J.S., 2001. Antioxidant activity of tocopherols, tocotrienols, and g-oryzanol components from rice bran against cholesterol oxidation accelerated by 2,20 -azobis(2-methylpropionamidine) dihydrochloride. Journal of Agricultural and Food Chemistry 49, 20772081. Yawadio, R., Tanimori, S., Morita, N., 2007. Identication of phenolic compounds isolated from pigmented rices and their aldose reductase inhibitory activities. Food Chemistry 101, 16441653. Ye, X., Al-Babili, S., Kloti, A., Zhang, J., Lucca, P., Beyer, P., Potrokus, I., 2000. Engineering the pro-vitamin A (beta-carotene) biosynthetic pathway into (carotenoid-free) rice endosperm. Science 287, 303305. Yokoyama, W., 2004. Nutritional properties of rice and rice bran. In: Champagne, E.T. (Ed.), Rice Chemistry and Technology. AACC, St. Paul, Minnesota, U.S.A., pp. 595609. Zhou, Z.K., Robards, K., Helliwell, S., Blanchard, C., 2004. The distribution of phenolic acids in rice. Food Chemistry 87, 401406.

Nucleotide polymorphisms in the waxy gene of NaN3-induced waxy rice mutants

Toong Long Jeng a, Chang Sheng Wang b, Tung Hai Tseng a, Min Tze Wu a, Jih Min Sung c, *
a
Division of Biotechnology, Taiwan Agricultural Research Institute, Wu Fong, Taichung 413, Taiwan, ROC Department of Agronomy, National Chung Hsing University, Taichung 402, Taiwan, ROC c Department of Food Science & Nutrition, Hungkuang University, Shalu, Taichung 433, Taiwan, ROC
b
Article history: Received 21 February 2008 Received in revised form 10 July 2008 Accepted 14 July 2008 Keywords: bp duplication G-to-T substitution Microsatellite Polymorphism Rice Waxy mutation
a b s t r a c t
Spontaneous and induced waxy phenotype, associated with endosperm containing little or no amylose, has been recognized in rice (Oryza sativa L.). Mutation of a dominant gene Wx into a recessive gene wx, which causes the inactivation or absence of granule bound starch synthase, is believed to be responsible for the change in endosperm starch leading to the waxy grain. In the present study, the nucleotide polymorphism in the Wx gene of rice genotype Tainung 67 (wild type) and its 35 NaN3-induced wx mutants were examined. Iodine staining conrmed that all the mutants had waxy grain trait. The G-to-T single base substitution analysis indicated that the wild type genotype Tainung 67 and its waxy mutants carried Wxb allele. Moreover, 23-bp duplication in exon 2 was detected in all the waxy mutants. Microsatellite polymorphism (CT)n was also detectable on the Wx gene of the tested genotypes and mutants, with at least 5 classes of (CT)n microsatellites identied at the Wx locus. Electrophoretic analyses also conrmed the observed nucleotide polymorphsim. Thus, nucleotide polymorphsim exist among NaN3-induced waxy mutants in rice. However, only the 23-bp duplication in exon 2 may be used as a molecular marker to characterize waxy grain trait in rice genotypes. 2008 Elsevier Ltd. All rights reserved.
1. Introduction Starch is the most abundant storage reserves in rice (Oryza sativa L.) grains. The starch reserves of rice grains generally contain 1730% amylose, which plays an important role in palatability and processing qualities (Ramesh et al., 1999; Zhou et al., 2002). Granule bound starch synthase (GBSS) is the key enzyme responsible for amylose synthesis in rice grain (James et al., 2003; Vandeputte and Delcour, 2004). Six loci are involved in the amylose synthesis of rice grains, with the dominant Wx gene that encodes GBSS being the major locus (Itoch et al., 2003). Mutation of the dominant Wx locus into recessive wx causes inactivation or absence of GBSS (Han et al., 2004), and subsequently results in rice grain of opaque endosperm that contains little or no amylose (Wang and Wang, 2002). Waxy (wx) mutants have been observed in both japonica and indica rice genotypes by spontaneous and induced mutation (Han et al., 2004; Hori et al., 2007; Sato and Nishio, 2003). Several functional molecular markers associated with Wx gene are detectable in waxy rice genotypes. An alternative splicing at the 50
* Correspondence author. Tel.: 886 4 26318652x5015. E-mail address: sungjm@sunrise.hk.edu.tw (J.M. Sung). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.07.009
splice site of intron 1 caused by a single base substitution (AGGT to AGTT) generally exists in waxy rice genotypes (Han et al., 2004; Isshiki et al., 1998; Itoch et al., 2003; Yamanaka et al., 2004). A duplication of 23-bp in the second exon, which causes the inactivation or loss of GBSS, is detectable only in waxy rice genotypes (Hori et al., 2007; Jeng et al., 2007; Wanchana et al., 2003). Polymorphic cytosine and thymidine (CT)n microsatellites in the 50 untranslated region are also noted in waxy rice genotypes, but with the number of alleles (4 alleles) consistently identied much less than non-waxy rice genotypes (10 alleles) (Bao et al., 2006). Many gamma ray- or chemical mutagen-induced waxy mutant lines have been obtained in rice, and the point mutations in Wx gene have been identied (Isshiki et al., 2001; Sato and Nishio, 2003). In this study, DNA polymorphisms of the Wx gene among Tainung 67 and its NaN3-induced wx mutant lines (Wang et al., 2002) were investigated. The G-to-T single base substitution and 23-bp duplication, which were generally used to differentiate waxy and non-waxy rice varieties on a molecular level, were used to identify the possible variations in the Wx locus of the tested genotypes and mutants. CT-microsatellite polymorphism was also employed to compare the possible differences among the tested mutants. The results of this study are useful for further study in identifying and breeding the waxy rice genotype with desirable agronomic traits.
T.L. Jeng et al. / Journal of Cereal Science 49 (2009) 112116
113
2. Materials and methods A total of 38 rice (O. sativa L.) accessions, including wild type genotype Tainung 67 and its 35 NaN3-induced waxy mutants, and 2 naturally mutated waxy genotypes TKW 1 and TCSW 1, were obtained from the Taiwan Agricultural Research Institute (Table 1). Additionally, a NaN3-induced non-waxy recovery of SA 419 (a NaN3induced waxy mutant of genotype Tainung 67) was also used to identify the possible effect of 23-bp duplication on the expression of wx gene. Seeds were sown in the nursery plots. The 3-leaf stage seedlings were transplanted to experimental plots (3 6 m) at a hill spacing of 30 15 cm with 3 seedlings per hill. Each plot received a basal application of fertilizer before transplanting (24 kg N ha1, 36 kg P2O5 ha1 and 24 kg K2O ha1) and 3 top-dressings of fertilizer at 20 d (6 kg N ha1, 9 kg P2O5 ha1 and 6 kg K2O ha1), 40 d (9 kg N ha1, 13.5 kg P2O5 ha1 and 9 kg K2O ha1) and 60 d (9 kg N ha1, 13.5 kg P2O5 ha1 and 9 kg K2O ha1) after transplanting (Jeng et al., 2003). All plants were grown to maturity. Mature dry seeds were hand-cut with a razor blade, and the halfseeds without embryo were soaked in distilled water at 4 C overnight. After the treatment, the half-seeds were sliced and stained with potassium iodide for visual examination (Itoch et al., 2003). Genomic DNA was extracted from leaves of 30-day-old seedlings of each genotype or line (Doyle and Doyle, 1990). For sequence analysis of the Wx alleles, a fragment was amplied by PCR (Invitrogen Co., USA) using two primer pairs (i.e., 50 -CAAGCTGGAAA AGCAAAAG-30 and 50 -TTGACCAACTCGGCTACTAA, 50 -ATGTCTCTCG CCACTGGA-30 and 50 -CTCAGCCACAACGCTGGTAT-30 ). PCR amplication was carried out by using 10 ng genomic DNA, 2 mM of each primer and 1 U rTag DNA polymerase (TaKaRa Co., Japan) at 56.4 C with 32 cycles. DNA sequencing was performed on an ABI 373 automated sequencer following the manufacturers instruction (Applied Biosystems, Inc., USA) using Dye Terminator Cycle Sequence Ready Reaction Kit (Perkin Elmer Co., USA). The G-to-T substitution at the 50 leader intron splice donor site of the Wx alleles was examined using a derived cleaved amplied polymorphic sequence (dCAPS) technique developed by Yamanaka et al. (2004). The forward 50 -TGTTGTTCATCAGGAAGAACATCTC CAAG-30 and reverse primers 50 -TTAATTTCCAGCCCAACACC-30 generate a unique EcoT14I restriction site characteristic of the Wxa allele. Total DNA was isolated from leaf, and the DNA fragments at the splice donor site of the rst intron were amplied by PCR (Invitrogen Co., USA). PCR amplication was carried out by using 100 ng genomic DNA, 0.4 mM of each primer and 1 U rTag DNA polymerase at 56 C with 32 cycles. Five microliters of each PCR product was digested with EcoT14I in a total volume of 20 mL at 37 C overnight. After digestion, each digest was electrophoresed in a 30 g kg1 Nusieve 3:1 agarose gel with ethidium bromide staining. The 23-bp duplication was detected using the technique detailed by Wanchana et al. (2003). The forward 50 -TGCAGA GATCTTCCACAGCA-30 and reverse primers 50 -GCTGGTCGTCAC GCTGAG-30 were used to generate amplicons specic for Wx alleles.
The forward primer was marked by IRDye700 (MWG-Biotech Co., Germany). A reaction mixture (20 mL) containing 20 ng of genomic DNA as the template, 0.2 mM of dNTPs, 0.2 mM of each primer, 0.5 U of Taq DNA polymerase, 50 mM HCl, 2.0 mM MgCl2, and 10 mM Tris-HCl (pH 8.3), and deionized H2O were added to make 20 mL nal volume. The amplication reactions were carried out by the following prole: 94 C pre-denaturation for 2 min followed by 32 cycles of (94 C denaturation for 30 s, 60 C annealing for 30 s and 72 C for 1 min) and the nal extension at 72 C for 5 min. The PCR products were then electrophoresed in 70 g kg1 polyacrylamide gel with LICOR 4300 DNA Analyzer (Licor Co., Germany). Similar PCR conditions were also used to detect CT-microsatellite polymorphism by using the forward 50 -CTTTGTCTATCTCAAGACAC-30 and reverse primers 50 -TTGCAGATGTTCTTCCTGATG-30 to generate amplicons specic for Wx alleles. 3. Results and discussion Notable differences in appearance of de-hulled grains were observed among the tested genotypes and NaN3-induced mutants (Fig. 1). All the tested mutants had chalky endosperm. Of these samples, the majority of mutants had white pericarp. Only 2 mutants had dark blue color pericarp (SA 526 and SA 572) and 3 mutants had brown color pericarp (SA 897, SA 674 and SA 1399). In Southeast Asia, waxy rice is grown mainly as a staple food (Wanchana et al., 2003). In Taiwan, waxy rice, particularly the waxy rice with colored pericarp, is generally reserved for use in festival foods and desserts (Jeng et al., 2007). Jeng et al. (2006) previously reported that, under adverse environmental conditions, the grain yield of NaN3-induced low amylose mutant SA 419 (code number 1 in Table 1 and Figs. 13) was better than its wild type genotype Tainung 67, and therefore was recommended to waxy rice growers and food manufacturers. In this regard, these 5 mutants with dark-
Table 1 The 35 NaN3-induced rice mutants of genotype Tainung 67 Code 1 2 3 4 5 6 7 8 9 Mutant SA 419 SA 402 SA 420 SA 472 SA 893 SA 897 SA 1917 SA 1170 SA 1445 Code 10 11 12 13 14 15 16 17 18 Mutant SA 1472 SA 491 SA 492 SA 495 SA 497 SA 502 SA 523 SA 526 SA 543 Code 19 20 21 22 23 24 25 26 27 Mutant SA 572 SA 573 SA 580 SA 588 SA 649 SA 674 SA 860 SA 861 SA 908 Code 28 29 30 31 32 33 34 35 Mutant SA SA SA SA SA SA SA SA 967 1047 1200 1220 1314 1324 1398 1399
The code numbers are used to identify the tested mutants presented on Figs. 13.
Fig. 1. The de-hulled grains of rice genotype Tainung 67 (TNG67) and its 35 NaN3induced mutants as well as two naturally occurred waxy mutants TKW 1 (japonica type) and TCSW 1 (indica type). The numbers marked in the gure match the code numbers listed on Table 1.
114
T.L. Jeng et al. / Journal of Cereal Science 49 (2009) 112116
Fig. 2. The iodine-stained endosperms of rice genotype Tainung 67 (TNG67) and its 35 NaN3-induced mutants as well as two naturally occurred waxy mutants TKW 1 (japonica type) and TCSW 1 (indica type). The numbers marked in the gure match the code numbers listed on Table 1.
colored pericarp could also be recommended to growers and food manufacturers, if their palatability, processing qualities and grain yield were proved promising. Rice genotype is considered as waxy with 020 g kg1 of grain amylose (Jeng et al., 2006). The opaque endosperm is generally regarded as a phenotypic marker for waxy rice (Bao et al., 2006; Han et al., 2004). This grain trait can be easily detected using iodine staining, which is a very sensitive technique to detect amylose in various tissues (Itoch et al., 2003). In the present study, the dark blue color of iodine staining suggested that the Wx gene was expressed in cultivar Tainung 67 (200 g kg1 amylose) (Jeng et al., 2006) (Fig. 2). On the other hand, all the waxy mutants (genotype TCSW 1, genotype TKW 1 and 35 NaN3-induced mutants) had orange color of stained endosperm, suggesting that they either have no or severely reduced amylose (Fig. 2). Several nucleotide polymorphisms are associated with the Wx gene, namely single nucleotide polymorphism (G/T) in the intron 1,
23-bp duplication in the exon 2 and CT-microsatellites (Bao et al., 2002; Wanchana et al., 2003; Yamanaka et al., 2004). Extensive research has indicated that two functional Wx alleles exist in cultivated rice genotypes, with Wxa having AGGT sequence and Wxb having an alternative AGTT sequence at 50 splice site in intron 1 (Han et al., 2004; Isshiki et al., 1998; Itoch et al., 2003; Yamanaka et al., 2004). A mutation from AGGT sequence to AGTT sequence at 50 splice site of Wx intron 1 leads to incomplete post-transcriptional processing of Wx pre-mRNA. Thus, the expression of Wxa is higher than Wxb at RNA, GBSS and amylose levels (Isshiki et al., 1998). Waxy rice genotypes generally have no detectable levels of spliced mRNA as a result of this mutation (Wang et al., 1995). They are characterized as the Wxb allele based on the G-to-T base substitution (Wang et al., 1995) and have very low or no amylose in mature grains (Wang and Wang, 2002). Nevertheless, Yamanaka et al. (2004) recently examined 353 waxy rice strains collected from various regions of Asia, and found that the collected waxy strains had both Wxa (AGGT)- and Wxb (AGTT)-derived alleles. They further indicated that the Wxb (AGTT)-derived allele was predominant, and Wxa (AGGT) allele was detected in only a minority of samples. Thus, G-to-T base substitution alone might not be a reliable marker for differentiating waxy and non-waxy rice varieties at a molecular level. In the present study, nucleotide sequence was also analyzed in genotype Tainung 67 and its 9 NaN3-induced waxy mutants as well as 2 waxy genotypes TKW 1 (japonica type waxy rice) and TCSW 1 (indica type waxy rice) (Table 2). Two genotypes Nipponbare (japonica genotype with intermediate amylose level) and 9311 (indica genotype with high amylose level), in which their nucleotide sequence had been analyzed and released on internet http://www.ncbi.nlm.nih.gov/, were also added for comparison (Table 2). The results showed that only indica genotype 93-11 had AGGT sequence at putative 50 leader intron. All the other tested accessions had AGTT sequence, regardless of their origin (indica or japonica). These results suggest that all the tested waxy mutants carry the Wxb allele. On the other hand, both japonica type genotypes Tainung 67 and Nipponbare (both with intermediate amylose level) have AGTT sequence. These results seem to indicate that the G-to-T polymorphism might be a useful molecular marker for identifying the rice genotype with grain amylose below intermediate level, but it is not sufcient to identify the waxy starch grain trait.
Fig. 3. Nucleotide polymorphisms of (A) 23-bp duplication, (B) G-to-T substitution and (C) (CT)n repeats on Wx alleles of rice genotype Tainung 67 (TNG67) and its 35 NaN3-induced mutants as well as two naturally occurred waxy mutants TKW 1 (japonica type) and TCSW 1 (indica type). A high amylose genotype TN 1 is added for comparison (a is wild type genotype Tainung 67, b is high amylose genotype TN 1, c is genotype TCSW 1, d is genotype TKW 1, M is marker). The numbers marked in gures match the code numbers listed on Table 1.
T.L. Jeng et al. / Journal of Cereal Science 49 (2009) 112116 Table 2 Comparisons of nucleotide sequences of Wx alleles of rice genotype Tainung 67 (TNG67) and its NaN3-induced mutants .exon 1.intron 1.exon 2. 1.50.142.1389. TNG 67 Nipponbare 93-11 TKW 1 TCSW 1 SA 419 SA 402 SA 420 SA 491 SA 492 SA 893 SA 1917 SA 1170 SA 1445 .(CT)18.T.(ACGGGTTCCAGGGCCTCAAGCCC)1 .(CT)18.T.(ACGGGTTCCAGGGCCTCAAGCCC)1 .(CT)18.G.(ACGGGTTCCAGGGCCTCAAGCCC)1 .(CT)18.T.(ACGGGTTCCAGGGCCTCAAGCCC)2 .(CT)16.T.(ACGGGTTCCAGGGCCTCAAGCCC)2 .(CT)15.T.(ACGGGTTCCAGGGCCTCAAGCCC)2 .(CT)15.T.(ACGGGTTCCAGGGCCTCAAGCCC)2 .(CT)19.T.(ACGGGTTCCAGGGCCTCAAGCCC)2 .(CT)17.T.(ACGGGTTCCAGGGCCTCAAGCCC)2 .(CT)17.T.(ACGGGTTCCAGGGCCTCAAGCCC)2 .(CT)17.T.(ACGGGTTCCAGGGCCTCAAGCCC)2 .(CT)17.T.(ACGGGTTCCAGGGCCTCAAGCCC)2 .(CT)16.T.(ACGGGTTCCAGGGCCTCAAGCCC)2 .(CT)17.T.(ACGGGTTCCAGGGCCTCAAGCCC)2
115
Two naturally mutated waxy genotypes TKW 1 (japonica type) and TCSW 1 (indica type) were added for comparison. Additionally, the nucleotide sequences of genotypes Nipponbare (japonica type) and 93-11 (indica type) are collected from internet http://www.ncbi.nlm.nih.gov/.
While apparently required for the mutation, the presence of the single base substitution (G-to-T) at the 50 splice site of intron 1 alone does not ensure that waxy mutation is expressed (Prathepha and Baimai, 2004; Yamanaka et al., 2004). Wanchana et al. (2003) identied a unique duplication of 23-bp in the second exon, which was present only in the waxy rice genotypes. The 23-bp duplication in exon 2 would result in a frame shift and subsequently cause a premature stop codon in the second exon and bring about degradation of mRNA by nonsense-mediated decay (Isshiki et al., 2001). Hori et al. (2007) examined 29 waxy genotypes and a 23-bp duplication at the second exon was detected in 27 genotypes (including paddy and upland waxy genotypes), and the two other waxy genotypes were considered to have a mutation (insertion of 7764 bp in exon 9) different from the 23-bp duplication. We also analyzed TKW 1, TCSW 1 and 9 NaN3-induced waxy mutants derived from Tainung 67 and detected 23-bp duplication in all the tested waxy mutants (Table 2). As expected, no 23-bp insertion was detected on non-waxy japonica genotypes Tainung 67, Nipponbare and indica genotype 93-11. Thus, the ndings of Wanchana et al. (2003) are still valid even though the waxy mutation is induced by chemical mutagen NaN3, at least for these NaN3-induced waxy mutants derived from genotype Tainung 67. Polymorphic cytosine and thymidine (CT)n microsatellites are noted in the 50 -untranslated region of the Wx gene (Bergmaan et al., 2001; Prathepha and Baimai, 2004), which seems to correlate well with the various amylose contents in non-waxy genotypes (Bao et al., 2002). Bao et al. (2002) examined a set of waxy rice genotypes, and found 4 microsatellite alleles, (CT)16, (CT)17, (CT)18, (CT)19 at the wx locus, of which (CT)17 was the most frequent. A similar conclusion was also reported from other studies (Han et al., 2004; Prathepha and Baimai, 2004). In the present study, a total of 5 classes of (CT)n microsatellites, located at 55-bp upstream of the putative 50 leader intron splice-junction were detected. Among the tested waxy mutants, 5 waxy mutants carried (CT)17 allele, 2 mutants contained (CT)15 allele, 2 mutants contained (CT)16 allele, a mutant had a (CT)18 allele and a mutant had (CT)19 allele (Table 2). Similar (CT)18 allele were also detected on non-waxy genotypes TNG 67, Nipponbare and 93-11. It should be pointed out that the (CT)15 allele was the only allele not detected in other studies (Bao et al., 2002; Han et al., 2004; Prathepha and Baimai, 2004). CT repeats are present in the 50 -untranslated region and variations of the number of the CT repeats from 8 to 20 have been report in nonwaxy genotype (Olsen and Purugganan, 2002). Therefore, CT
repeats is not considered to be the cause of the waxy trait (Hori et al., 2007). However, CT repeats might potentially serve as a molecular marker for other starch qualities of waxy rice, such as G/T-related thermal properties as well as some of the pasting viscosity parameters (Bao et al., 2002). Direct nucleotide sequencing is labor intensive and costly to analyze a large sample set. Electrophoretic analysis is known to be an accurate technique to analyze large numbers of samples quickly, without the need for direct sequencing determination (Yamanaka et al., 2004). In the present study, all the 35 waxy mutants (including the accessions used for nucleotide sequence analysis) and their wild type genotype Tainung 67 plus 2 waxy genotypes TKW 1 and TCSW 1 that were differentiated by iodine staining (Fig. 2) were used to identify 23-bp duplication polymorphism through electrophoresis. A well-known non-waxy indica type genotype TN 1 (Taichung Native 1) was also added to serve as a comparison. The electrophoretic results (Fig. 3A) showed that no 23-bp duplication was detected in genotypes Tainung 67 and TN1, which contain intermediate and high grain amylose, respectively. As expected, all the waxy accessions (including cultivars TKW 1 and TCSW 1) had the 23-bp insertion in the second exon (Fig. 3A). This premature stop codon due to the 23-bp insersion would cause the loss of function of GBSS encoded by the Wx gene (Sato and Nishio, 2003). For further verifying the functional role of the 23-bp duplication, an electrophoresis on a NaN3-induced non-waxy recovery line of SA 419 was conducted and the results indicated that no 23-bp duplication at the second exon was found for this non-waxy recovery line of SA 419 (Fig. 4). This result clearly indicates that the 23-bp duplication at the second exon is essential for NaN3-induced waxy mutation, at least for these mutants derived from genotype Tainung 67. Electrophoresis of G-to-T polymorphism indicated that only indica genotype TN 1 having high grain amylose content had AGGT sequence on intron 1 splice site (Fig. 3B). On the other hand, all the tested waxy mutants, including TKW 1 and TCS 1, as well as their wild type TNG 67 contained the AGTT sequence at the 50 splice site of the rst intron (Fig. 3B). Moreover, variations in the length of CTmicrosatellite in the 50 -untranslated region of the wx gene (Bao et al., 2002; Prathepha and Baimai, 2004) was also obtainable in the tested accessions (Fig. 3C). In conclusion, the present results indicate that signicant differences in grain appearance are observed among the tested genotypes and NaN3-induced waxy mutants. Nucleotide polymorphisms also exist among the tested mutants. G/T polymorphism analysis indicates that all the NaN3-induced waxy mutants carry Wxb allele. However, the presence of the single base substitution (G-to-T) at the 50 splice site of intron 1 alone does not ensure that waxy mutation is expressed, because both wild type genotype Tainung 67 and non-waxy genotype Nipponbare (both are genotypes with intermediate grain amylose level) also shows same G/T substitution at 50 splice site of intron 1. Additionally, at least 5 classes of (CT)n microsatellites at the Wx locus are identied among the tested waxy mutants. However, the CT repeats are also not the cause of the waxy trait. On the other hand, a 23-bp
Fig. 4. Nucleotide polymorphisms of 23-bp duplication on Wx alleles of (1) mutant SA 419 and its (2) non-waxy recovery.
116
T.L. Jeng et al. / Journal of Cereal Science 49 (2009) 112116 Itoch, K., Ozaki, H., Okada, K., Hori, H., Takeda, Y., Mitsui, Y., 2003. Introduction of Wx transgene into rice wx mutants lead to both high- and low-amylose rice. Plant and Cell Physiology 44, 473480. James, M.G., Denyer, K., Meyers, A.M., 2003. Starch synthesis in the cereal endosperm. Current Opinions in Plant Biology 6, 215222. Jeng, T.L., Wang, C.S., Yseng, T.H., Sung, J.M., 2007. Expression of granule-bound starch synthase in developing rice grain. Journal of the Science of Food and Agriculture 87, 24562463. Jeng, T.L., Tseng, T.H., Wang, C.S., Chen, C.L., Sung, J.M., 2003. Starch biosynthesizing enzymes in developing grains of rice cultivar Tainung 67 and its sodium-azideinduced rice mutant. Field Crops Research 84, 261266. Jeng, T.L., Tseng, T.H., Wang, C.S., Chen, C.L., Sung, J.M., 2006. Yield and grain uniformity in contrasting rice genotypes suitable for different growth environment. Field Crops Research 99, 5966. Olsen, K.M., Purugganan, M.D., 2002. Molecular evidence on the origin and evolution of glutinuous rice. Genetics 162, 941950. Prathepha, P., Baimai, V., 2004. Variation of Wx microsatellite allele, waxy allele distribution and differentiation of chloroplast DNA i n a collection of Thai rice (Oryza sativa L. Euphytica 140, 231237. Ramesh, M., Ali, S.Z., Bhattacharya, K.R., 1999. Structure of rice starch and its relation to cooked-rice texture. Carbohydrate Polymers 38, 337347. Sato, Y., Nishio, T., 2003. Mutation detection in rice waxy mutants by PCR-RF-SSCP. Theoretical and Applied Genetics 107, 560567. Vandeputte, G.E., Delcour, J.A., 2004. From sucrose to starch granule to starch physical behaviour: a focus on rice starch. Carbohydrate Polymers 58, 245266. Wanchana, S., Toojinda, T., Tragoonrung, S., Vanavichit, A., 2003. Duplicated coding sequence in the waxy allele of tropical glutinuous rice (Oryza sativa L.). Plant Science 165, 11931199. Wang, Y.J., Wang, L., 2002. Structures of four waxy rice starches in relation to thermal, pasting and textural properties. Cereal Chemistry 79, 252256. Wang, C.-T., Tseng, T.-H., Lin, C.-Y., 2002. Rice biotech research at the Taiwan Agricultural Research Institute. Asia Pacic Biotechnology News 6, 950956. Wang, Z.-Y., Zheng, F.-Q., Shen, G.-Z., Gao, J.-P., Snutad, D.P., Li, M.-G., Zhang, J.-L., Hong, M.-M., 1995. The amylose content in rice endosperm is related to the post-transcriptional regulation of the waxy gene. Plant Journal 7, 613622. Yamanaka, S., Nakamura, I., Watanabe, K.N., Sato, Y.-I., 2004. Identication of SNPs in the waxy gene among glutinuous rice cultivars and their evolutionary signicance during the domestication process of rice. Theoretical and Applied Genetics 108, 12001204. Zhou, Z., Robards, K., Helliwell, S., Blanchard, C., 2002. Ageing of stored rice: changes in chemical and physical attributes. Journal of Cereal Science 35, 6578.
duplication in the second exon is present in all the NaN3-induced waxy mutants. This 23-bp duplication appears to be necessary for the waxy trait in genotype Tainung 67. Thus, the 23-bp duplication in exon 2 may be used as a molecular marker to characterize waxy grain trait in rice breeding program. Acknowledgements We gratefully acknowledge nancial support from the National Science Council of ROC. References
Bao, J.S., Corke, H., Sun, M., 2002. Microsatellites in starch-synthesizing genes in relation to starch physicochemical properties in waxy rice (Oryza sativa L.). Theoretical and Applied Genetics 105, 898905. Bao, J.S., Corke, H., Sun, M., 2006. Microsatellites, single nucleotide polymorphisms and a sequence tagged site in starch-synthesizing genes in relation to starch physicochemical properties in non-waxy rice (Oryza sativa L.). Theoretical and Applied Genetics 113, 11851196. Bergmaan, C.J., Delgado, J.T., McClung, A.M., Fjellstrom, R.G., 2001. An improved method for using a microsatellite sequence closely linked to the waxy gene to determine amylose class. Cereal Chemistry 78, 257260. Doyle, J.J., Doyle, J.L., 1990. Isolation of plant DNA from fresh tissue. Focus 12, 1315. Han, Y., Xu, M., Liu, X., Yan, C., Korban, S.S., Chen, X., Gu, M., 2004. Genes coding for starch branching enzymes are major contributors to starch viscosity characteristics in waxy rice (Oryza sativa L.). Plant Science 166, 357364. Hori, Y., Fujimoto, R., Sato, Y., Nishio, T., 2007. A novel wx mutation caused by insertion of a retrotransposon-like sequence in a glutinous cultivar of rice (Oryza sativa). Theoretical and Applied Genetics 115, 217224. Isshiki, M., Yamamoto, Y., Satoh, H., Shimamoto, K., 2001. Nonsense-mediated decay of mutant waxy mRNA inrice. Plant Physiology 125, 13881395. Isshiki, M., Morino, K., Nakajima, M., Okagaki, R.J., Wessier, S.R., Izawa, T., Shimamoto, K., 1998. A naturally occurring functional allele of the rice waxy locus has a GT to TT mutation at the 50 splice site of the rst intron. Plant Journal 15, 133138.

Amaranth (Amaranthus hypochondriacus) as an alternative crop for sustainable food production: Phenolic acids and avonoids with potential impact on its nutraceutical quality
A.P. Barba de la Rosa a, *, Inge S. Fomsgaard b, Bente Laursen b, Anne G. Mortensen b, L. Olvera-Martnez c, C. Silva-Sanchez a, A. Mendoza-Herrera d, J. Gonzalez-Castaneda e, A. De Leon-Rodrguez a
a
Institute for Scientic and Technological Research at San Luis Potosi, Molecular Biology Division, Camino a La Presa San Jose No. 2055, Lomas 4a seccion, San Luis Potos, SLP, Mexico Department of Integrated Pest Management, Faculty of Agricultural Sciences, Aarhus University, Denmark c CBETa 196 Villa de Pozos, San Luis Potos, SLP, Mexico d Horticultural Science Department, HFSB Room 202, 2133 TAMU, College Station, United States e Instituto de Ciencias Agrcolas, Universidad de Guanajuato, Irapuato, Gto, Mexico
b
Article history: Received 7 May 2008 Received in revised form 4 July 2008 Accepted 10 July 2008 Keywords: Crop yield LC/MS/MS Protein content Phytochemicals RAPD
a b s t r a c t
The demand for food is increasing, not only to meet food security for growing populations, but also to provide more nutritious food, rich in good quality proteins and nutraceutical compounds. The amaranth (Amaranthus hypochondriacus) plant, in addition to its high nutritive and nutraceutical characteristics, has excellent agronomic features. The objective of the present study was to analyze some physical and proximal-nutritional properties of amaranth seeds obtained from different varieties grown in arid zones and characterize their phenolic acids and avonoids. Two commercial (Tulyehualco and Nutrisol) and two new (DGETA and Gabriela) varieties of A. hypochondriacus were grown at the Mexican Highlands zone. Tulyehualco and DGETA varieties had higher seed yield of 1475 and 1422 kg ha1, respectively, comparable to corn and soybean production in agricultural areas. Gabriela had the highest protein content of 17.3%, but all varieties had an adequate balance of essential amino acids. Polyphenols as rutin (4.010.2 mg g1 our) and nicotiorin (7.24.8 mg g1 our) were detected. Amaranth can be cultivated in arid zones where commercial crops cannot be grown; the seeds besides their well known nutritive characteristics could be a source of phenolic compounds of high antioxidant properties. 2008 Elsevier Ltd. All rights reserved.
1. Introduction The challenge for agricultural practices to increase food production to obtain food security still persists after 40 years of the Green Revolution (Hobbs, 2007). The rst Millennium Development Goal is to reduce hunger and poverty by 2015 (Dixon et al., 2006). The demand for food is increasing, not only because of the growing population, but also to provide more nutritious food with high protein quality and nutraceutical compounds. Water resources, especially surface and ground water will be more limited as domestic and industrial needs increase just as it is limited in semi-desert zones with low precipitation, and so future crops must be more suited to low water use. Amaranth (Amaranthus hypochondriacus) is a crop naturally resistant to water decit and is a good source of nutritious seeds; the seeds have high
* Corresponding author. Tel.: 52 4448342000; fax: 52 4448342010. E-mail address: apbarba@ipicyt.edu.mx (A.P. Barba de la Rosa). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.07.012
amounts of protein containing essential amino acid such as lysine and methionine; also signicant levels of squalene, an important precursor for all steroids (He et al., 2002), were reported in amaranth oil. Since the beginning of the 1980s, amaranth has been rediscovered and several reports have tried to promote it as a basic crop (Kauffman, 1992). In addition to nutritional characteristics, amaranth plants have agronomic features identifying it as an alternative crop where cereals and vegetables cannot be grown (dry soils, high altitudes and high temperatures) (Omami et al., 2006). In general, the selection of promising genotype in a breeding program is based on various criteria, with the most important being nal crop yield and seed quality (Kozak et al., 2008). The aim of the research presented in this paper was to analyze four amaranth varieties and select the most promising for growing in arid zones like the Mexican Highlands or arid zones of Southern Europe. Variety selection was carried out on the basis of seed yield and proximal quality. The presence of a range of phytochemical compounds that contribute to nutraceutical properties of the seeds was determined as well.
118
A.P. Barba de la Rosa et al. / Journal of Cereal Science 49 (2009) 117121
2. Materials and methods 2.1. Plant material Two commercial varieties (Tulyehualco and Nutrisol) and two new varieties (DGETA and Gabriela) of A. hypochondriacus were cultivated at the CBTa 196 Villa de Pozos in San Luis Potosi, under low watering conditions, precipitations between 25 and 35 mm. The experiment was in randomized complete blocks with ve blocks per variety. Seeds were drilled manually in four rows, each 6 m long, 80 cm apart and with 30 cm spacing in the row (Henderson et al., 2000). Plants were watered at 35% of eld capacity, at the beginning of soil preparation (April 2002) and in August 2002. After stem elongation was complete (November 2002), hand-harvest was carried out at maturity by cutting plants at soil level. 2.2. Physical analysis Measurements of morphologic parameters were taken from 10 randomly chosen plants. Plant height was measured from the ground to the top of the inorescence head. Spike length was measured from the base to the top of inorescence. For seed yield, all plants from each block were hand-harvest, seeds were cleaned and weighted, and yield was calculated and extrapolated to kg ha1. A sample of 100 seeds per triplicate was weighted and extrapolated to obtain the weight of 1000 seeds. The seeds were milled to obtain the our able to pass through a 100-mesh screen and our yield was determined as the percentage of our recovery. The ours were defatted with hexane at a our/hexane ratio of 1:10 (w/v) under continuous stirring for 4 h at 4 C, the slurry was centrifuged at 9000 g for 20 min and the our was air-dried at room temperature and stored at 4 C (Barba de la Rosa et al., 1992). 2.3. Chemical analysis Total nitrogen (method 954.01), fat (920.39), ash (923.03), crude ber (962.09), and moisture (925.09) contents of the ours were determined according to AOAC (1990). Nitrogen was determined with a Kjeltec system (Tecator, Sweden). Protein was calculated from total nitrogen using a factor of 5.85 (Barba de la Rosa et al., 1992). Fat amount was obtained from 4 h hexane extraction. Ash was calculated from the weight remaining after heating the sample at 550 C for 2 h. Moisture measurement was determined from sample weight loss after oven drying at 110 C for 4 h. All samples were analyzed in triplicate. 2.4. Essential amino acid analysis Amino acid analysis of amaranth ours was performed by reversed-phase high performance liquid chromatography (RP-HPLC) using a HPLC Waters 600 with a uorescence detector 2475 (Waters) and performed by AccQ-Tag amino acid analysis kit and following the manuals instructions (Waters). Samples of 5 mg of amaranth ours were hydrolyzed with 200 mL of 6 N HCl for 24 h at 110 C in vacuo system; one or two phenol crystals were added as an oxygen scavenger (Barba de la Rosa et al., 1992). The samples were neutralized with 1.2 N NaOH and dehydrated. Subsequently, samples were derivatized with 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate. A 5 mL sample was applied to the column AccQ-Tag C18 (Waters). All samples were analyzed in triplicate. 2.5. Phytochemical analysis The freeze-dried amaranth our samples were crushed and homogenized before phytochemicals were extracted using an
Accelerated Solvent Extraction 200 system (Dionex) (ASE), following the protocol reported by Carlsen et al. (2008). Briey, 0.1 g of the freeze-dried and homogenized sample was transferred to the 33 mL extraction cells followed by 5-g of chemically inert Ottawa sand (2030 mesh particle size, Fisher Chemicals) after heating at 400 C. A lter was placed on top of the sample, and the extraction cell was lled with glass balls that had been heated at 400 C. The eluent was 70% methanol, 30% water (v/v). The protocol for the ASE extraction was the following: preheat for 5 min, heat for 5 min, static for 3 min, ush 80%, purge for 60 s, 4 cycles, pressure 1500 psi, and temperature 80 C. Phytochemical extracts were collected in vials, which were lled with eluent to maintain 46.1 g ( 50 mL) weight for all extracts, and stored at 20 C until chemical analysis. The extracts were ltered on a Sartorius SRP 15 0.45-m lter (PTFE membrane) and diluted with water in a ratio of 1:1. Phytochemical analysis was carried out using an Applied Biosystems MDS Sciex API3200 Liquid chromatography-Ion Trap Quadrupole Mass Spectrometer (LC/MS/MS) with turbo electrospray ionization in a negative multiple reaction monitoring (MRM) mode. Pure reference compounds were used for identication of the phenolic acids and polyphenols based on a comparison of fragmentation pattern and retention times. Standard curves for commercially obtained pure standards were prepared in 50% methanol and 50% water (v/v) at the following concentrations: 3.13, 6.25, 12.5, 25.0, 50.0, 100, 200, 400, and 800-mg L1. The standard curve was used for the quantitative determination. The analyzed compounds were three polyphenols, rutin (CAS no. 153-18-4, Extrasynthese, France), isoquercitrin (CAS no. 482-35-9, Fluka, Germany), and nicotiorin (CAS no. 17650-84-9, Carl Roth, Germany), and three phenolic acids, vanillic acid (CAS no. 121-34-6, Fluka, Germany), 4-hydroxybenzoic acid (CAS no. 99-96-7, Fluka, Germany), and syringic acid (CAS no. 530-57-4, Sigma, Germany). 2.6. Electrophoretic patterns Protein fractions were obtained by following the procedure described elsewhere with some modications (Barba de la Rosa et al., 1992). Briey, for albumin fraction extraction, a suspension of our/distilled water (1:10 w/v) was prepared and stirred for 1 h at 4 C and centrifuged at 9000 g for 20 min at 4 C. The resulting pellet was re-suspended in 10 mM Na2HPO4, pH 7.5, 1 mM EDTA, 0.1 M NaCl, stirred and centrifuged as before, the supernatant was named as globulins 7S. The pellet was again re-suspended in phosphate buffer but containing 0.8 M NaCl; the soluble fraction was named globulins 11S. The glutelin fraction was extracted with 0.1 M NaOH. The protein patterns were analyzed by SDS-PAGE at 15% w/v polyacrylamide in a Mini-Protean III system (Bio-Rad). Samples of 1 mg mL1 were dissolved in Laemmli sample buffer (Laemmli, 1970). Electrophoresis was conducted at a constant current of 20 mA per gel for 23 h. After electrophoresis, the gel was stained with Coomassie Brilliant Blue G250 at a nal concentration of 0.25%. Destaining was achieved by washing the gel for 24 h with acetic acid/methanol/water (4.5:4.5:1 v/v/v). 2.7. RAPD analysis DNA of the four cultivars was extracted from fresh young leaves by a standard procedure. DNA quantication was by measurement of absorbance at 260/280 nm. The extracted DNA was dissolved in double distilled water (Dellaporta et al., 1983). For amplication, the pair of oligonucleotides denominated as 23: 50 -CCC GCC TCC C-30 and 43: 50 -AAA ACC GGG C-30 were used as reported by Chan and Sun (1997) and following the protocol described by those authors: 3 mM MgCl2, 0.1% Triton X-100, 75 ng primer, and 30 ng genomic DNA was used in 25 mL reaction mix. PCR amplication
A.P. Barba de la Rosa et al. / Journal of Cereal Science 49 (2009) 117121 Table 1 Physical properties of amaranth seeds Variety Plant height (cm) 159.0a 154.2b 149.1b 161.3a Spike length (cm) 68.0a 54.5a 28.2b nd Seed yield (kg ha1) 1475a 1422a 1204b 1121b Weight of Flour yield 1000 seeds (%) (g) 0.6a 0.7a 0.7a 0.6a 87.3a 89.1a 87.6a 88.2a
119
Table 3 Essential amino acid composition of different amaranth variety ours (g of amino acid 100 g1 of crude protein) Amino acid Ile Leu Lys Met Cysa Phe Tyrb Thr Val
a b
Tulyehualco 3.2 5.8 6.9 0.3 7.9 3.8 3.4
DGETA 2.2 5.0 5.2 1.7 3.7 3.3 2.6
Gabriela 3.1 6.5 7.5 6.5 8.3 4.8 3.6
Nutrisol 3.5 6.6 6.7 7.2 7.3 4.6 3.9
FAO/WHO/UNU (1986) 1.3 1.9 1.6 1.7 1.9 0.9 1.3
Tulyehualco DGETA Gabriela Nutrisol
Means of three replicates within in the same column (physical property) with different letter are signicantly different at P 0.05.
Requirements for methionine cysteine. Requirements for phenylalanine tyrosine.
was performed in a iCycler (Bio-Rad) using the following cycle prole: 1 cycle at 94 C for 2 min followed by 45 cycles at 94 C for 1 min, 35 C for 2 min, and 72 C for 2 min; and the nal cycle at 72 C for 7 min. The amplication products were electrophoresed on 1.4% agarose gels and visualized using ethidium bromide staining. Negative controls were routinely used to check for possible contamination. For RAPD analysis, the banding patterns were recorded in a GelDoc photodocumentator (Bio-Rad) and gels analyzed with Quantity-one software (Bio-Rad). The bands with the same molecular weight and mobility were treated as identical fragments; the presence of a band was coded as 1, and absence coded as 0. The data matrices were analyzed with the Squared Euclidean nearest neighbor method (Statgraphics Plus v5.0 software, Statistical Graphics Corp). Dendrograms were produced from the results. 2.8. Statistical analysis Statistical analysis was performed and Tukeys test was used to determine signicance of differences among means. Signicance was when compared means differed at P 0.05 (Reyes-Castaneda, 1983). 3. Results and discussion 3.1. Physical properties of amaranth seeds The physical properties of amaranth plant and seeds are shown in Table 1. Tulyehualco and Nutrisol plants were the highest (161.3 and 159.0 cm, respectively), but Tulyehualco had the longest spike (68 cm). Tulyehualco and DGETA had the highest seed yields of 1475 and 1422 kg ha1, respectively. When grown in drier environments, amaranth seeds yields of 1050 and 410 kg ha1 were reported (Henderson et al., 2000). The amaranth seed yields obtained in our work compare favorably to commercially important crops like soybean (1532 kg ha1) and maize (2315 kg ha1) (SAGARPA, 2002). There were no differences in the weight of 1000 seeds among varieties. Flour yield is an important quality parameter of wheat, and since our is an alternative use of amaranth seeds, this parameter was also measured. There were no differences among the varieties with a mean value of 88% (Table 1), values that are similar to wheat our yields (Paredes-Lopez et al., 1990). In Central Mexico, amaranth grain is paid at double the price of maize
grain. In this sense, farmers could have double benets, seeds of high nutritional quality and higher incomes for the rural families (Islas-Gutierrez and Islas-Gutierrez, 2001). 3.2. Proximal composition of amaranth Amaranth protein content (Table 2) was higher than values reported for cereals (Paredes-Lopez et al., 1990). The variety Gabriela had the highest seed protein content (17.3%) followed by Tulyehualco, Nutrisol and DGETA (15.0%, 15.3%, and 14.8%, respectively). The fat contents were similar among the varieties; however, at higher proportions (79%), than for cereals such as wheat (2.1%) and maize (4.5%). There were no differences in ash contents between varieties, but Gabriela had the highest ber content, 2.5% (Table 2). Amaranth has been considered as a new source of ber and methods for fractionation to get high-ber amaranth products have been reported (Tosi et al., 2001). The essential amino acids in amaranth seeds (Table 3) are ideal according to FAO requirements for adults (FAO/WHO/UNU, 1986). 3.3. Phytochemicals in amaranth seed ours It has been reported that amaranth seed our contains polyphenols (avonoids) with relatively high antioxidant status. For this reason, amaranth has been recommended for use in balanced diets (Gorinstein et al., 2007). However, to the best of our knowledge, no results have ever been published before on the types of polyphenols present in amaranth seed ours. In this study, three polyphenols (rutin, isoquercitrin, and nicotiorin), were identied and quantied (Table 4). Rutin was present at higher concentrations (10.1 mg g1 our) in Tulyehualco seed our, while the highest amount of nicotiorin (7.2 mg g1 our) was found in the Gabriela variety (Table 4). Rutin has been reported to be present in amaranth leaves (Suryavanshi et al., 2007), but this is the rst time it has been reported in seed ours. Polyphenols in which sugar groups are beta-linked, as in the case of the three polyphenols found in this study, are easily degraded in the intestine of humans and animals due to the abundance of beta-glucosidase enzyme that liberates the aglyconic
Table 2 Proximal composition of amaranth seeds (% dry basis) Variety Tulyehualco DGETA Gabriela Nutrisol Protein* 15.0b 14.8b/c 17.3a 15.3b Lipids 8.1a 7.9a/b 8.9a 8.6a Ash 3.3a 3.3a 3.9a 3.5a Fiber 2.2a 1.9b 2.5a 2.0a/b
Table 4 Phenolic acids and avonoids present in different amaranth variety our (mg metabolite g1 our)a Metabolite Isoquercitrin Nicotiorin Rutin 4-Hydroxybenzoic acid Syringic acid Vanillic acid
a
Tulyehualco 0.5 5.5 10.1 1.7 0.8 1.8
DGETA 0.5 5.6 5.8 2.0 0.7 1.7
Gabriela 0.3 7.2 4.0 2.2 nd 1.8
Nutrisol nd 4.8 4.7 1.9 nd 1.5
*N 5.85. Means of triplicates in the same column with different letter are signicantly different at P 0.05.
Mean of two replicates, nd not detected.
120
Fig. 1. Electrophoretic patterns of amaranth proteins fractions. A) albumins, B) 7S globulins, C) 11S globulins, D) glutelins. Lane S molecular weight standard; lane 1 Tulyehualco; lane 2 DGETA; lane 3 Gabriela; lane 4 Nutrisol. The arrow shows the differential band found in glutelin fraction.
moiety of the molecules. Several health effects have been published, which generally concern the uptake of the aglyconic groups, quercetin and kaempferol (Donovan et al., 2007). Rutin and its metabolites may effectively modulate advanced glycation end product (AGE) formation, which is associated with numerous pathologies. Clinical diseases possibly accelerated by AGEs include neuropathy, nephropathy, retinopathy, joint stiffness, senile cataracts, Alzheimers disease, and cardiovascular disease (CervantesLaurean et al., 2006). Polyphenols such as quercetin have been shown to serve as a protective defense against oxidative damage in vivo (Meyers et al., 2008). Nicotiorin has been claimed to have protective effects on reducing memory dysfunction (Huang et al., 2007); recent results have proved a strong pharmacological basis for its potential therapeutic role in cerebral ischemic illness (Li et al., 2006). Phenolic compounds were proved to have antioxidant activity, where rutin, vanillic acid, ferulic acid, and quercetin have been reported in rice wine samples and correlated with antioxidant activities (Que et al., 2006). In relation to phenolic acids, one paper was published in which several phenolic acids were identied and quantied in amaranth seeds using HPLC-UV analysis (Klimczak et al., 2002). In our study, using LC/MS/MS, three phenolic acids; vanillic acid; 4-hydrozybenzoic acid, and syringic acid, were identied and quantied (Table 4). The levels of syringic acid and vanillic acid are similar to those reported in rye where syringic acid has been related with their bitterness (Heinio et al., 2008).
fraction is rich in peptides of angiotensin converting enzyme inhibitor (Silva-Sanchez et al., 2008). The amaranth glutelin pattern (Fig. 1D) had high similarity with 11S globulins. There are three main polypeptides groups, one at 2225 kDa, the second at 3538 kDa, and the third around 55 kDa. In this fraction, there was an additional band at 65 kDa in the Tulyehualco and DGETA varieties; these varieties were also those with the higher seed yields. Abugouch et al. (2003) suggested that glutelins are globulin-like proteins with hexameric structure of approximately 300 kDa, and it is known that the amaranth globulin sequence has high homology with the Cupin domain that has been associated with resistance to extremes of environment, as heat, sulfur nutrition, water stress (Higashi et al., 2006; Khuri et al., 2001). Former studies by LC/MS/MS analysis showed that the glutelin fraction is rich in antihypertensive, and also contains the anticarcinogenic lunasin-like peptide (Silva-Sanchez et al., 2008). 3.5. RADP analysis Morphological variations cause confusion or misidentication of varieties of the same species so that genetic analyses could be necessary for correct identication. In comparative studies using RAPD, RFLP and/or allozyme markers, the most valuable tool for correct identication of genetic variations was RAPD (Chan and Sun, 1997). Four different varieties of amaranth were studied in this work; the amplied fragments were analyzed and were grouped, generating a dendogram (Fig. 2). DGETA and Gabriela are the closer varieties and probably both derived from Nutrisol. The data indicate
3.4. Electrophoretic patterns of seed storage amaranth proteins Electrophoretic patterns of the different protein fractions in amaranth seeds were obtained under denaturing conditions. The albumin fraction was the main fraction in amaranth proteins; all varieties had a characteristic band at 34 kDa (Fig. 1A) as reported by Barba de la Rosa et al. (1992). The albumin fraction was characterized by a group of molecular weight around 18 kDa; this group of proteins is known as the MRPs (methionine rich proteins), proteins that contain 1618% of methionine (Segura-Nieto et al., 1994). Another group of proteins in the albumins fraction was located at 4080 kDa. Albumins are comparable with egg-white proteins, and this protein fraction has been used as an egg substitute to prepare bread (Silva-Sanchez et al., 2004). There are two groups of globulin seed storage proteins, the 7S globulins were extracted with 0.1 M NaCl (Fig. 1B) and had a main band around 38 kDa. Globulin fractions extracted with 0.8 M NaCl (Fig. 1C) showed the characteristic pattern of 11S-like globulins, the group of acid polypeptides (3538 kDa) and the group of basic polypeptides (2225 kDa). The band of high molecular weight (55 kDa) was previously reported as a globulin precursor (Barba de la Rosa et al., 1992). Recently we reported that the 11S globulin
Fig. 2. Dendrograms of Amaranth hypochondriacus varieties based on RAPDs analysis using the Squared Euclidean nearest neighbor method from Statgraphics software. Lane 1 Tulyehualco; lane 2 DGETA; lane 3 Gabriela; lane 4 Nutrisol; lane 5 control.
121
that Tulyehualco variety is the one most distantly related to the other three. 4. Conclusions As interest in amaranth cultivation has increased and breeders have produced a large number of new varieties adapted to different zones (Gimplinger et al., 2008; Henderson et al., 2000). However, some of these new varieties are only new names for old varieties. RAPD analysis showed that the old variety, Tulyehualco, could be the progenitor of the other three varieties used in this work; this variety also had the highest seed yield and has a good balance of amino acids. Here, for the rst time, some secondary metabolites are reported to be present in amaranth seed ours. These metabolites, namely rutin, isoquercitrin, and nicotiorin are compounds that could have implications in the prevention of several illnesses. The highest rutin content was found in Tulyehualco, while nicotiorin was found in highest concentration in the Gabriela variety. Tulyehualco, because of the high seed yield, high protein and rutin content, seems to be an excellent amaranth variety to grow in arid areas worldwide. The knowledge of amaranth as a source of phytochemicals will increase their importance as a potential source of those compounds in the human diet. Amaranth therefore has great potential as a sustainable crop that could improve household food security and farm incomes. In arid areas where commercial crops such as beans, maize or rice cannot be grown, amaranth cultivation could contribute to the rst Millennium Development Goal of reducing hunger and poverty. Acknowledgments Thanks are due AG Alpuche Solis for his help in plant harvesting, and to Fundacion Produce San Luis Potosi for nancial support. This work was partially supported nancially by the European Commission 6th Framework Programme, AMARANTH:FUTURE FOOD, Contract No. 032263. Thanks to Dagmar Janovska for reviewing the manuscript. References
Abugouch, L.E., Martinez, E.N., Anon, M.C., 2003. Inuence of the extracting solvent upon the structural properties of amaranth (Amaranthus hypochondriacus) glutelin. Journal of Agricultural and Food Chemistry 51, 40604065. AOAC, 1990. Ofcial Methods of Analysis, 15th ed. Association of Ofcial Analytical Chemists, Washington, DC. Barba de la Rosa, A.P., Gueguen, J., Paredes-Lopez, O., Viroben, G., 1992. Fractionation procedures, electrophoretic characterization and amino acid composition of amaranth seed protein. Journal of Agricultural and Food Chemistry 40, 931936. Carlsen, S.C.K., Mortensen, A.G., Oleszek, W., Piacente, S., Stochmal, A., Fomsgaard, I.S., 2008. Variation in Flavonoids in Leaves, Stems and Flowers of White Clover Cultivars. Natural Products Communications 3, 12991306. Cervantes-Laurean, D., Schramm, D.D., Jacobson, E.L., Halaweish, I., Bruckner, G.G., Biossonneault, G.A., 2006. Inhibition of advanced glycation end product formation on collagen by rutin and its metabolites. The Journal of Nutritional Biochemistry 17, 531540. Chan, K.F., Sun, M., 1997. Genetic diversity and relationships detected by isozyme and RAPD analysis of crop and wild species of Amaranthus. Theoretical and Applied Genetics 95, 865873. Dellaporta, S.L., Wood, J., Hicks, J.B., 1983. A plant DNA minipreparation. Plant Molecular Biology Reporter 1, 1922. Dixon, J., Nalley, L., Kosina, P., La Rovere, R., Hellin, J., Aquino, P., 2006. Adoption and economic impact of improved wheat varieties in the developing world. Journal of Agricultural Science 144, 489502. Donovan, J.L., Manach, C., Faulks, R.M., Kroon, P.A., 2007. Absorption and metabolism of dietary plant secondary metabolites. In: Crozier, A., Clifford, M.N., Ashihara, H. (Eds.), Plant Secondary Metabolites: Occurrence, Structure and Role in the Human Diet. Blackwell Publishing, London, pp. 303351.
FAO/WHO/UNU, 1986. Special report. Energy and protein requirements. Cereal Foods World 3, 694695. Gimplinger, D.M., Schulte aufm Erley, G., Dobos, G., Kaul, H.P., 2008. Optimum crop densities for potential yield and harvestable yield of grain amaranth are conicting. European Journal of Agronomy 28, 119125. Gorinstein, S., Medina-Vargas, O.J., Jaramillo, N.O., Arnao-Salas, I., MartnezAyala, A.L., Arancibia-Avila, P., Toledo, F., Katrich, E., Trakhtenberg, S., 2007. The total polyphenols and the antioxidant potentials of some selected cereals and pseudocereals. European Food Research and Technology 225, 321328. He, H.P., Cai, Y., Sun, M., Corke, H., 2002. Extraction and purication of squelene from amaranthus grain. Journal of Agricultural and Food Chemistry 50, 368372. Heinio, R.-L., Liukkonen, K.-H., Myllymaki, O., Pihlava, J.-M., Adlercreutz, H., Heinonen, S.-M., Poutanen, K., 2008. Quantities of phenolic compounds and their impacts on the perceived avor attributes of rye grain. Journal of Cereal Science 47, 566575. Henderson, T.L., Joson, B.L., Schneiter, A.A., 2000. Row spacing, plant population, and cultivar effects on grain amaranth in the northern great plains. Agronomy Journal 92, 329336. Higashi, Y., Hirai, M.Y., Fujiwara, T., Naito, S., Noji, M., Saito, K., 2006. Proteomic and transcriptomic analysis of Arabidopsis seeds: molecular evidence for successive processing of seed proteins and its implication in the stress response to sulfur nutrition. Plant Journal 48, 557571. Hobbs, P.R., 2007. Conservation agriculture: what is it and why is it important for future sustainable food production. Journal of Agricultural Science 145, 127137. Huang, J.L., Fu, S.T., Jiang, Y.Y., Cao, Y.B., Guo, M.L., Wang, Y., Xu, Z., 2007. Protective effects of Nicotiorin on reducing memory dysfunction, energy metabolism failure and oxidative stress in multi-infarct dementia model rats. Pharmacology, Biochemistry and Behavior 86, 741748. Islas-Gutierrez, J., Islas-Gutierrez, F., 2001. Rentabilidad de los cultivos de amaranto y maz para la zona central de Mexico. Agricultura Tecnica de Mexico 27, 143151. Kauffman, C., 1992. The status of grain amaranth for the 1990s. Food Reviews International 8, 165192. Khuri, S., Bakker, F.T., Dunwell, J.M., 2001. Phylogeny, function and evolution of the Cupins, a structurally conserved, functionally diverse superfamily of proteins. Molecular Biology and Evolution 18, 593605. Klimczak, I., Malecka, M., Pacholek, B., 2002. Antioxidant activity of ethanolic extracts of amaranth seeds. Nahrung-Food 46, 184186. Kozak, M., Bocianowski, J., Rybinski, W., 2008. Selection of promising genotypes based on path and cluster analyses. Journal of Agricultural Science 146, 8592. Laemmli, U.K., 1970. Cleavage of structural protein during the assembly of head of bacteriophage T4. Nature 227, 680685. Li, R.P., Guo, M.L., Zhang, G., Xu, X.F., Li, Q., 2006. Neuroprotection of nicotiorin in permanent focal cerebral ischemia and in neuronal cultures. Biological and Pharmaceutical Bulletin 29, 18681872. Meyers, K.J., Rudolf, J.L., Mitchell, A.E., 2008. Inuence of dietary quercetin on glutathione redox status in mice. Journal of Agricultural and Food Chemistry 56, 830836. Omami, E.N., Hammes, P.S., Robbertse, P.J., 2006. Differences in salinity tolerance for growth and water-use efciency in some amaranth (Amaranthus spp.) genotypes. New Zealand Journal of Crop and Horticultural Science 34, 1122. Paredes-Lopez, O., Barba de la Rosa, A.P., Hernandez-Lopez, D., Carabez-Trejo, A., 1990. Amaranto: caractersticas y aprovechamiento agroindustrial. Programa Regional de Desarrollo Cientco y Tecnologico, Washington, DC. Que, F., Mao, L., Pan, X., 2006. Antioxidant activities of ve chinese rice wines and the involvement of phenolic compounds. Food Research International 39, 581587. Reyes-Castaneda, P. (Ed.), 1983. Bioestadstica aplicada, Agronoma, Biologa, Qumica. Trillas, Mexico, DF. SAGARPA, 2002. Sistema de informacion agropecuaria de consulta. SIACON, Mexico, DF. Segura-Nieto, M., Barba de la Rosa, A.P., Paredes-Lopez, O., 1994. Biochemistry of amaranth proteins. In: Paredes-Lopez, O. (Ed.), Amaranth: Biology, Chemistry, and Technology. CRC Press, Boca Raton, FL, pp. 75106. Silva-Sanchez, C., Barba de la Rosa, A.P., Leon-GalvanF, de Lumen, B.O., De Leon Rodrguez, A., Gonzalez de Meja, E., 2008. Bioactive peptides in amaranth (Amaranthus hypochondriacus) seed storage proteins. Journal of Agricultural and Food Chemistry 56, 12331240. Silva-Sanchez, C., Gonzalez-Castaneda, J., De Leon Rodrguez, A., Barba de la Rosa, A.P., 2004. Functional and rheological properties of amaranth albumins extracted from two mexican varieties. Plant Foods for Human Nutrition 59, 169174. Suryavanshi, V.L., Sathe, P.A., Baing, M.M., Singh, G.R., Lakshmi, S.N., 2007. Determination of Rutin in Amaranthus spinosus Linn. Whole Plant Powder by HPTLC. Chromatography 65, 767769. Tosi, E.A., Re, E., Lucero, H., Masciarelli, R., 2001. Dietary ber obtained from amaranth (Amaranthus cruentus) grain by differential milling. Food Chemistry 73, 441443.

Relationship of milled grain percentages and owering-related traits in rice

Rodante E. Tabien a, *, Stanley Omar P.B. Samonte a, Emmanuel R. Tiongco b
a b
Texas A&M University System, AgriLife Research and Extension Center, 1509 Aggie Drive, Beaumont, TX 77713, USA Philippine Rice Research Institute, Science City of Munoz, Nueva Ecija 3119, Philippines
Article history: Received 29 December 2007 Received in revised form 12 July 2008 Accepted 15 July 2008 Keywords: Rice Flowering duration Milling qualities Flowering traits
a b s t r a c t
The economic value of harvested rice is determined by the grain yield and the percentages of head rice (at least 3/4 the length of a head or kernel) and total milled rice. This study was conducted to determine the effects of owering-related traits such as duration of owering, rate of owering, heading, and duration from heading to maturity on head rice and total milling percentages of rice. Flowering data, gathered for two years from 105 long grain rice genotypes grown in Beaumont, Texas were analyzed for their effects on and relationship with milling traits. A positive linear relationship was obtained for rate of owering and the duration from heading to maturity but negative for duration of owering and days to heading. Genotypes with early heading had relatively shorter owering durations, and genotypes with shorter owering duration had higher head rice and total milled rice. A faster rate in attaining 100% owering and more days from heading to maturity were favorable in increasing head rice and total milled grains. The duration from the start of owering to heading or to 100% owering can be used in the evaluation and selection for high head rice and total milled rice percentages in rice. Published by Elsevier Ltd.
1. Introduction The economic value of harvested rice is determined by the grain yield and the percentages of head rice (at least 3/4 the length of a head or kernel) and total milled rice. The high value of head rice relative to broken grains inuences the cultivar choice and the preand post-harvest handling practices to maximize farm income. These milling traits had shown signicant positive direct effects on gross income, thus these should be considered in selecting cultivars to plant and in developing a high income-grossing genotype (Samonte et al., 2006). Milling traits are inuenced by the environment, cultural practices, milling processes, drying, and genotype. Environmental factors reported to affect head rice percentages include meteorological conditions such as relative humidity, air temperature, and rainfall (Banaszek and Siebenmorgen, 1990; Jodari and Linscombe, 1996; Thompson and Mutters, 2006). Cultural practices include time of draining (Counce et al., 1990; McCauley and Way, 2002), harvesting (McCauley and Way, 2002) and nitrogen fertilization (Jongkaewwattana et al., 1993; Leesawatwong et al., 2005; Seetanum and De Datta, 1973). Milling processes that inuence grain milling traits include the type of mill (Bautista and Siebenmorgen, 2002), milling cylinder speeds (Dilday, 1989), degree of milling (Reid et al., 1998), temperature and and duration (Cnossen et al.,
* Corresponding author. Tel.: 1 409 752 2741. E-mail address: retabien@ag.tamu.edu (R.E. Tabien). 0733-5210/$ see front matter Published by Elsevier Ltd. doi:10.1016/j.jcs.2008.07.015
2003; Schluterman and Siebenmorgen, 2007), cooling methods and milling procedures (Pan et al., 2005), and head rice separation methods (Lloyd et al., 2001). Drying, as it relates to grain moisture content, is critical in getting high percentage of head rice (Banaszek and Siebenmorgen, 1990; Dilday, 1989; Siebenmorgen and Jindal, 1986; Siebenmorgen et al., 1992). Grain moisture content (MC) of 1518% has been found ideal for milling grain rice (Jodari and Linscombe, 1996), but the most recent study indicated that the optimal harvest moisture content (HMC) at which milling traits peaked for long grain cultivar was 1822% while medium grains needed 2224% HMC (Siebenmorgen et al., 2007). Cultivars differed signicantly in their milled rice percentages (Dilday, 1989). Genotypic traits reported to affect head rice include tillering and kernel weights at lower seeding rates (Gravois and Helms, 1996), kernel thickness (Lu and Siebenmorgen, 1995), panicle type and grain weight (Wang et al., 2007), and panicle length and maturity (Jongkaewwattana and Geng, 2002). Nitrogen fertilization has been shown to increase head rice percentage (Jongkaewwattana et al., 1993), but this increase was dependent on genotype (Perez et al., 1996). The optimum N rates and harvest grain moisture to obtain the maximum head rice percentage of different cultivars were also determined by Leesawatwong et al. (2005). Siebenmorgen et al. (2007) found that the magnitude of head rice reduction at different moisture content was dependent on cultivar. In estimating the number of days from emergence to heading (50% of the panicles are owering) for each breeding line, rice breeders have observed that owering duration from the onset to
R.E. Tabien et al. / Journal of Cereal Science 49 (2009) 122127
123
100% owering and maturity vary across lines. Chau and Kunze (1982) reported large variation in grain maturity and consequently MC within a eld (1343%). There was a large MC variation in individual kernels in panicle and in plant, and among plants (Bautista and Siebenmorgen, 2005; Holloway et al., 1995; Kocher et al., 1990). Differences in grain weights and quality within panicle were found to be dependent on variety (Cheng et al., 2007; Liu et al., 2005). Rice owers asynchronously. The owers at the top of the panicles open rst and the bottom orets open last, and the owering duration may last up to 15 days (Holloway et al., 1995; Luh and Luh, 1991). These variations have resulted in multi-modal MC distribution among rice kernels and to a variation in amylose content (Cheng et al., 2007). Grain structure was also found to be related to grain lling which had an impact on head rice (Jongkaewwattana and Geng, 2001). Decreasing kernel dimensions, volume and density, and amylose content from top to lower part of the panicle was observed in cultivars with different maturities (Cheng et al., 2007; Jongkaewwattana and Geng, 2002) and this non-uniformity inuenced head rice recovery. It is hypothesized that owering-related traits such as duration of owering, rate of owering, heading, and duration from heading to harvest can affect head rice and total milling percentages. Hence, this study determined these relationships. 2. Experiment The Uniform Regional Rice Nursery (URRN) located at the Texas A&M University System, AgriLife Research and Extension Center, Beaumont, TX (29 570 N, 94 300 W), is part of a multi-state yield trials. The nursery is composed of elite genotypes from various rice breeding programs in the USA, and it has 200 entries per year. The 105 long grain genotypes that were common across the 2005 and 2006 URRN were used in this study. The soil in the research center is Beaumont, an Entic Pelludert (ne, montmorillonitic, and thermic). The planting dates of the breeding materials in both years were mid-April. Each cropping year, the eld experiments were fertilized with 225 kg N ha1 equally split into three and applied at planting, one month after planting, and at panicle differentiation. The entries were laid out in randomized complete block design (RCBD) with two to four replications for entries at the preliminary and the advanced trials, respectively, but only two replications were used for data gathering in this study. Each plot had six 3-m long rows spaced at 25 cm. The eld data obtained on a plot basis for each line from two replications included: days from emergence to onset (start), 25, 50 (heading), 75, and 100% owering, and days to harvest. Milling samples were obtained at maturity from rice grown in a 1-sq m area in each plot. These were threshed and dried right after harvest. The head rice and total milled rice percentages were obtained following the standard procedures (Fan et al., 2000).
To produce brown rice, the 125-g rough rice samples were hulled by a single pass through a rubber roll huller (Satake Engineering Co., LTD., Tokyo, Japan). The brown rice was then milled for 54 s using a McGill No. 2 mill (Rapsilver Supply Co Inc., Brookshire, TX). The generated milled rice was separated into head rice and broken fractions with a shaker/separator (#12 screen 4.76 mm; Seedburo Equipment Co., Chicago, IL); The weight of the grain recovered was used to calculate the percentage of total milled rice and head rice based on 125 g of rough rice. From eld data, the following owering durations from seedling emergence were estimated on a plot basis: start of owering (onset) to 50% owering (FD050), 50% owering to 100% owering (FD50100), start of owering to 100% owering (FD0100), and rate of owering (slope of the number of days to 25, 50, 75, and 100%). Date of harvest was used in estimating the number of days from 50% owering to maturity. Analysis of variance was used to evaluate the effect of genotype, year and its interaction to days from emergence to 50% owering (heading), duration of owering (FD050, FD50100, FD0100), rate of owering, days from heading to maturity, and milling traits (percent head rice and percent total milled rice). Regression and correlation analyses were conducted to determine the relationships between owering-related traits (owering duration, rate of owering, heading, days from heading to maturity) and the percentages of head rice and total milled rice, and the days to heading with owering duration. Using the relationships derived, estimated values for the owering-related traits at 70 and 55% total and head rice percentages, respectively, were generated. 3. Results and discussion 3.1. Year, genotype, and genotype year effects on owering and milling traits In breeding trials across years or locations, the number of days from emergence to heading was signicantly affected by genotype and genotype year (Table 1). The number of days to heading ranged from 82.0 days (recorded in line RU0401084, cvs. Spring and Trenasse (known as very early rice cultivars), to 96.5 days (line RU0503126) in 2005 and 72.0 days (cv. Trenasse) to 90.0 days (line RU0501139) in 2006. Mean number of days to heading was 88.1 days in 2005 and 79.9 days in 2006. Although it was much shorter for the genotypes to have 50% owering in 2006, the variations were not enough to indicate signicance between the two years. Earlier genotypes or later genotypes in 2005 showed the same owering response in 2006 indicating a strong genetic control for this trait. A QTL in chromosome 8 that explains more than 94.9% of variation in heading date was reported recently (Zhang et al., 2006). The same locus was identied in several different populations (Lin et al., 2003; Xiong et al., 1999; Zhuang et al., 1997); thus this QTL could be present in some of the 105 genotypes evaluated in this study.
Table 1 Means squares from the analysis of variance for owering-related traits of 105 long grain genotypes grown at Beaumont, Texas in 2005 and 2006 Source of variation DF Days to heading Duration 050% owering 0.09 85.95 9.75 1.84** 1.04 1.16 Duration 50100% owering 0.68 0.68 3.62 0.68** 0.52* 0.37 Duration 0100% owering 0.38 103.09 24.38 3.24** 1.66 1.50 Flowering rate Duration from heading to maturity 21.94 237.75 60.95 47.92** 13.51** 1.60 Whole milled rice percentage 0.92 5637.87* 11.55 72.48** 36.58** 8.48 Total milled rice percentage 0.19 1323.77* 0.97 17.65** 6.31** 1.20
Replication Year (Y) Error Genotype (G) Y G interaction Error b
1 1 1 104 104 208
21.94 7109.49 60.95 36.11** 4.72** 1.60
13.30 889.76 341.23 32.24** 18.17 16.86
*, ** Signicant at the 5 and 1% level of probability, respectively.
124
Genotype signicantly affected FD050, FD50100, and FD0100, while genotype year signicantly affected FD50100 (Table 1). These indicate that there was enough variation in the duration of owering among 105 genotypes. Genotype FD050 (averaged across years) ranged from 2.0 days (line RU0401145 and cv. Spring) to 5.5 days (line RU0503098), FD0100 ranged from 3.5 days (cv. Spring) to 9.5 days (line RU0503098), while FD50100 ranged from 1.0 days (cv. Trenasse) to 4.0 days (line RU0503098). Mean FD050, FD0100, and FD50100 were 3.3, 5.5, and 2.2 days, respectively. Shorter duration of owering (050%, 50100%, or 0100%) was mostly obtained on very early season varieties like cv. Spring and Trenasse. Duration of owering, however, was not taken in yield trials, unlike heading date or the number of days from emergence to 50% owering. The highly signicant variation obtained and interaction noted suggests that these traits may be important in obtaining high quantity and quality harvests. Long owering duration may increase variation in kernels at harvest. Since the rst oret will be at a different stage relative to the late owering orets, there is likely to be high variability in size, degree of grain lling, and moisture content. Reports indicated signicant variation in kernel qualities within the panicle, plant or hill, and elds (Bautista et al., 2007; Jongkaewwattana and Geng, 2002), and these were related to head rice recovery. The rate of owering (% days1) was signicantly affected by genotype but not by year, and by the genotype year interaction (Table 1). The rate of owering across years ranged from 9.9% days1 (line RU0503098) to 25.08% days1 (line RU0502177) in 2005 and 10.81% days1 (line RU0503098) to 29.33% days1 (line RU0203032) in 2006. The mean across genotypes for the rate of owering was 17.65% days1 in 2005 and 20.56% days1in 2006. The average rate of owering in two years was slowest in RU0503098 (10.36 % days1) and fastest in cultivar Spring (25.66 % days1). Among the released cultivars, early maturing cultivars, such as Spring and Trenasse, the hybrid XP 723, and the newly released Presidio were consistently fast to reach 100% owering in two years. The wide variation observed in two years among genotypes suggests that some genotypes had much faster or slower rate to attain 100% owering than the rest of the entries. Similar to duration of owering, the rate of owering may affect the uniformity of the grains. The grains will be more uniform for fast owering genotypes than for slow owering genotypes. The number of days from heading to maturity was signicantly affected by genotype, and by genotype year interaction, but not by year (Table 1). In 2005, the shortest duration from heading to harvest was 29 days (line RU0503126) and the longest was 47 days (line RU0503181). In 2006, the shortest duration was 26 days (line RU0401179) and the longest was 49 days (line RU0501102). Average duration to harvest per year was 38 days in 2005 and 40 days in 2006. Among the released cultivars, Cocodrie, Trenasse, and hybrid XP 723 matured longest at 45, 44 and 42 days, respectively. The shortest duration among cultivars was from Banks at 34 days followed by Saber and Spring at 35 days. The variation among genotypes again reects potential differences in the duration of grain lling that can affect milling traits. Genotype, genotype year, and year had highly signicant effect on both head rice and total milled rice percentages (Table 1). Head rice percentages (averaged across years) ranged from 37.1% (line RU0501096) to 64.3% (line RU0504198) in 2005 and from 42.2% (cv. Priscilla) to 67.5% (cv. Cybonnet) in 2006. Mean (averaged across genotypes) head rice percentage was 51.5% in 2005 and 58.9% in 2006. Total milled rice percentages (averaged across years) ranged from 63.0% (line RU0301188) to 72.9% (cv. XP723) in 2005, and from 63.0% (line RU0501139) to 76.9% (line RU0502068) in 2006. Mean (averaged across genotypes) total milled rice percentage was 68.0% in 2005 and 71.5% in 2006. The results clearly indicate that better milling traits were obtained in 2006 compared
to 2005, and this can be due to differences in environmental factors such as temperature, rainfall, and relative humidity (Banaszek and Siebenmorgen, 1990; Cooper et al., 2006; Jodari and Linscombe, 1996; Thompson and Mutters, 2006). The reported signicant interaction of year and genotype (Gravois et al., 1991) was considered an important source of variation for head rice.
3.2. Days to heading and milling percentages Regression analysis shown in Fig. 1 revealed a linear relationship of days to heading with head rice and total milled rice. Although the amount of variation that can be explained by this trait was low in 2005 and much higher at 3038% in 2006, the same negative and highly signicant relationships were obtained in both years. The head rice and total milled rice percentages decreased with longer duration to heading. A reduction of 0.740.80% in head rice and 0.350.44% for total milled rice is possible per day increase in days to heading. Based on the 2006 linear relationship, a desirable total milled rice percentage of 70% and head rice percentage of 55% should have 83.44 days and 85.23 days to heading, respectively. Longer periods than these estimates will mean lower milling percentages. Very early heading genotypes are usually not selected in most rice breeding programs. In most cases, plants that head too early had fewer tillers and lighter biomass, thereby reducing grain yield (Xiao et al., 1998); these also had prolonged vegetative stage resulting in poor grain lling and lower grain yield. Grain yield and days to heading were positively correlated (data not shown); thus there should be a compromise to obtain both high yield and high total and head rice percentages. Using the two year data, correlation of heading to total and head rice percentages was very high, with r-values of 0.70 and 0.75, respectively. This indicates that heading may be useful for the indirect selection of high head rice and total milled rice percentages. However, one must consider its interaction with year and the potential impact of the environment. Actual use of this trait in selection will verify its effectiveness in improving milling traits. The overlapping of values for total and head rice as shown in Fig. 1, for instance, indicates variability if this trait will be used as selection index. These variations can be due to timing at which the 50% heading has occurred. Using historical data of two cultivars, average daily low temperature (high night time temperature) at 50% heading was shown to signicantly affect head rice (Cooper et al., 2006). High night time temperature was reported to decrease
Fig. 1. Relationship of number of days from emergence to heading, and head and total milled rice percentages in 105 long grain genotypes grown at Beaumont, Texas in 2005 and 2006.
125
grain dimensions, increase number of chalky grains, lower amylose content, and increase length of amylopectin chains, thereby affecting milling qualities (Cooper et al., 2008; Counce et al., 2005). 3.3. Flowering duration and milling percentages Flowering duration to 50% and 100%, and 50 to 100% owering were negatively correlated with milling traits (Table 2). The highest correlation was found between FD0100 and percentage head grain (r 0.464) followed by FD050 and head rice percentage (r 0.433). The FD50100 was also signicantly and negatively correlated with head rice and total milled rice percentages, but these correlation coefcients were lower than both FD050 and FD0100. These relationships indicate that the duration to reach 50% or 100% owering is important in obtaining high total and head rice percentages than the duration from 50100%. Moreover, these indicate the need for shorter duration of owering to insure better total and head grain percentages. The negative relationship between owering duration and milled rice percentages is shown in Fig. 2. There was a 2.4% and 1.1% decrease in head rice and total milled, respectively for every one day increase in ower duration. Grains from the rice panicle or from the entire plant, develop and mature asynchronously. At harvest, the distribution of MC of grains is multi-modal (Bautista and Siebenmorgen, 2005; Kocher et al., 1990). The range of grain MC decreases as the MC at harvest decreases suggesting that the ranges of MC are directly related to the maturity of the grains (Bautista and Siebenmorgen, 2005). These ranges of variation will be more at prolonged owering duration since the maturity of the grains will be different. The differences in owering result in wider variation in grain development, and the nal weight and quality of the grains (Wang et al., 2006). Any non-uniformity expressed in variable grain size and shape, grain lling, and maturity has negative effects on rice milling traits (Jongkaewwattana and Geng, 2001). Rice breeders commonly estimate the number of days from emergence to heading of their breeding lines. Adding days to start of owering to the list of evaluation criteria will enable the estimation of FD050 for indirect selection criterion for relatively higher head rice and total milled rice percentage. The mean of 105 genotypes for FD050 and FD0100, were 3.3 and 5.5 days, respectively, and the mean head rice and total milled rice percentages (averaged across years and genotypes) were 55.2 and 69.8%, respectively. Based on the regression equations obtained, the FD050 and FD0100 should be at less than 5.5 days in order to produce higher than average head rice and total milled rice percentages. Unaffected by year and the genotype year interaction, the trait of owering duration owering can be used in indirect selection relative to the number of days to heading. Although the correlation was lower compared to heading date, this trait can be a more reliable indicator of milling potentials. Actual use in the breeding program will prove its efciency in improving milling traits.
Fig. 2. Relationship of owering duration and whole and total milled rice percentages in 105 long grain genotypes grown at Beaumont, Texas in 2005 and 2006.
3.4. Rate of owering and milling percentages The rate of owering, similar to the duration of owering, was linearly and positively related to both head rice and total milled rice percentages as shown in Fig. 3. Increasing rate of owering was favorable in improving milling traits. This result indicates that a genotype should reach 100% owering at a faster rate in order to produce high quality grain for milling. The relationship of rate of owering and milling can be explained by the uniformity of grains at harvest. If the genotype has a fast rate of owering, the grains will be lled nearly at the same time and will be in a near-uniform stage during harvest. Synchronous grain development related to grain lling and maturity has a positive impact on milling qualities (Jongkaewwattana and Geng, 2001). Correlation of owering rate to milling percentages was signicant with r-values comparable to that of duration of owering, thus indicating the formers potential for indirect selection. Estimation of owering rate, however, needs several data points compared to the simpler gathering of data for duration of owering.
3.5. Duration of heading to maturity and milling percentages The number of days from heading to maturity did not have a signicant relationship with head rice percentages in 2005 and 2006. However, there were signicant linear relationships between
Table 2 Correlation between owering-related traits and milling percentages of 105 long grain rice genotypes grown at Beaumont, Texas in 2005 and 2006 Flowering duration Correlation coefcient (r) Whole milled rice Start of owering to 50% owering 50100% owering Start to 100% owering Heading date Flowering rate Duration for heading to maturity ** Signicant at the 1% probability level. 0.43** 0.27** 0.46** 0.70** 0.42** 0.18** Total milled rice 0.43** 0.190** 0.42** 0.75** 0.40** 0.43**
Fig. 3. Relationship of owering rate and whole and total milled rice percentages in 105 long grain genotypes grown at Beaumont, Texas in 2005.
126
the number of days from heading to maturity and total milled rice percentages in 2005 and 2006 (Fig. 4). For each day that the number of days from heading to maturity increased, total milled rice percentage increased by 0.24% in 2005 and 0.29% in 2006. The increase in percent total grain at longer duration from owering to harvest can be attributed to better grain lling. It was observed that grain lling affected grain traits such as weight and density, shape and size, which, in turn, affected milling traits (Siebenmorgen et al., 1992). Moreover, at 40 days after owering, most of the grains had 20% moisture content, and maximum total and head rice grain was obtained at 1222%. An opposite relationship was noted in Californian cultivars with different maturities. There was a quadratic relationship between maturity and head rice grain, and the percentage head rice increased from early and intermediate to late and very late maturing cultivars (Jongkaewwattana and Geng, 2002). This difference can be attributed to grain shape and size. Most of the Californian cultivars were medium grain, but this study focused on long grains that are popular in the Gulf Coast. Variation in grain size and shape contributed to the variation in rate and duration of grain lling (Jongkaewwattana and Geng, 2001) and ultimately milling quality. It was reported that late-heading entries had larger panicle, and the date of heading was found to be correlated with the number of grains per panicle (0.52) and spikelets per panicle (0.42) (Zhang et al., 2006). These results further support increased grain variations in late heading genotypes. Number of days to heading and maturity are important for rice adaptation to certain cultivation areas and to cropping seasons in most rice growing countries. Shorter maturity or early owering is needed in rainfed and upland areas and during the rainy season. However, a relatively longer maturity is needed to have higher grain yield in irrigated areas during the dry season planting. Therefore, it is important to consider the number of days to heading and maturity in selecting for higher total and head rice percentages, and high grain yield.
Fig. 5. Relationship of number of days from emergence to start of owering and the owering duration in 105 long grain genotypes grown at Beaumont, Texas in 2005 and 2006.
3.6. Duration of owering and the number of days from emergence to 50% owering The number of days from emergence to heading had a signicant linear relationship with owering duration (Fig. 5). The number of days from the start of owering to 50% was longer at prolonged duration to heading (50% owering). Across 2005 and 2006, the daily owering duration increased by 0.12 days as the number of days from emergence to heading increased. This
relationship suggests that late owering genotypes should be avoided to have genotypes with shorter duration of owering. Early owering genotypes should give better total and head grain percentages considering the positive relationship of heading date and the duration of owering, and the negative relationship of milling traits and the duration of owering. Breeders are not focusing on very early genotypes because of lower grain yields but since earliness is related to duration of owering and the duration of owering is correlated with milling traits, a trade-off is necessary. Based on the linear relationships obtained, a line should ower between 80 and 85 days. The US-released varieties included in the trials had 7286 days to 50% owering and the majority of cultivars were within the estimated range (data not shown). Based on these results, owering-related traits such as days to heading, rate of owering, duration of owering, and duration from heading to maturity can impact head rice and total milled rice percentages. These milling traits were positively related with rate of owering, and duration from heading to maturity but negatively related with duration of owering (050%, 50100% and 0100%) and days to heading. Rice genotypes with early heading had relatively shorter owering durations, and genotypes with shorter owering duration had higher head rice and total milled rice percentages. Faster duration to attain 100% owering and more days from heading to maturity were favorable in increasing percentage head rice and total grain. Estimating the duration from start of owering to heading or to 100% owering can be potential selection criteria in the indirect evaluation and selection of breeding lines for high head rice and total milled rice percentages. The number of days from heading to maturity can be considered in indirectly selecting for high total milled rice percentage. Acknowledgements The authors would like to acknowledge the nancial support from Texas Rice Research Foundation and the eld and laboratory assistance, especially the harvesting and milling of samples, provided by Chersty Harper, Patrick Frank, Joel Pace, Richard Boyd, and Pat Carre. References
Banaszek, M.M., Siebenmorgen, T.J., 1990. Head rice yield reduction rates caused by moisture adsorption. Transactions of the ASAE 33, 12631269. Bautista, R.C., Siebenmorgen, T.J., 2005. Individual kernel rice moisture content variability trends. Applied Engineering in Agriculture 21 (4), 637643.
Fig. 4. Relationship of number of days from heading and maturity on total milled rice percentages in 105 long grain genotypes grown at Beaumont, Texas in 2005 and 2006.
R.E. Tabien et al. / Journal of Cereal Science 49 (2009) 122127 Bautista, R.C., Siebenmorgen, T.J., 2002. Evaluation of laboratory mills for milling small samples of rice. Applied Engineering in Agriculture 18 (5), 577583. Bautista, R.C., Siebenmorgen, T.J., Counce, P.A., 2007. Rice kernel dimensional variability trends. Applied Engineering in Agriculture 23 (2), 207217. Chau, N.N., Kunze, O.R., 1982. Moisture content variation among harvested rice grains. Transactions of the ASAE 25, 10371040. Cheng, F.M., Zhang, Q.F., Zhu, H.J., Zhao, N.C., Wang, F., Chen, K.M., Zhang, G.P., 2007. The difference in amylose content within a panicle as affected by the panicle morphology of rice cultivars. Journal of Cereal Science 46, 4957. Cnossen, A.G., Jimenez, M.J., Siebenmorgen, T.J., 2003. Rice ssuring response to high drying and tempering temperatures. Journal of Food Engineering 59, 6169. Cooper, N.T.W., Siebenmorgen, T.J., Counce, P.A., Meullenet, J.-F., 2006. Explaining rice milling quality variation using a historical weather data analysis. Cereal Chemistry 83 (4), 447450. Cooper, N.T.W., Siebenmorgen, T.J., Counce, P.A., 2008. Effects of night time temperature during kernel development on rice physicochemical properties. Cereal Chemistry 83 (4), 447450. Counce, P.A., Siebenmorgen, T.J., Vories, E.D., Pitts, D.J.,1990. Time of draining and harvest effects on rice grain yield and quality. Journal of Production Agriculture 3, 436445. Counce, P.A., Bryant, R.J., Bergman, C.J., Bautista, R.C., Wang, Y.-J., Siebenmorgen, T.J., Moldenhauer, K.A.K., Meullenet, J.-F.C., 2005. Rice milling quality, grain dimensions, and starch branching as affected by high night temperatures. Cereal Chemistry 82 (6), 645648. Dilday, R.H., 1989. Milling quality of rice: cylinder speed vs. grain-moisture content at harvest. Crop Science 29, 15321535. Fan, J., Siebenmorgen, T.J., Yang, W., 2000. A study of head rice yield reduction of long and medium grain rice varieties in relation to various harvest and drying conditions. Transactions of the ASAE 43, 17091714. Gravois, K.A., Moldenhauer, K.A.K., Rohman, P.C., 1991. Genotypic and genotype environment effects for rough rice and head rice yields. Crop Science 31, 907911. Gravois, K.A., Helms, R.S., 1996. Seeding rate effects on rough rice yield, head rice, and total milled rice. Agronomy Journal 88, 8284. Holloway, G.E., Siebenmorgen, T.J., Counce, P.A., Lu, R., 1995. Causes of multi-modal moisture content frequency distribution among rice kernels. Applied Engineering in Agriculture 11 (4), 561565. Jodari, F., Linscombe, S.D., 1996. Grain ssuring and milling yields of rice cultivars as inuenced by environmental conditions. Crop Science 36, 14961502. Jongkaewwattana, S., Geng, S., 2001. Inter-relationships amongst grain characteristics, grain-lling parameters and rice (Oryza sativa L.) milling quality. Journal of Agronomy and Crop Science 187, 223229. Jongkaewwattana, S., Geng, S., 2002. Non-uniformity of grain characteristics and milling quality of California rice (Oryza sativa L.) of different maturities. Journal of Agronomy and Crop Science 188, 161167. Jongkaewwattana, S., Geng, S., Brandon, D.M., Hill, J.E., 1993. Effect of nitrogen and harvest grain moisture on head rice yield. Agronomy Journal 85, 11431146. Kocher, M.F., Siebenmorgen, T.J., Norman, R.J., Wells, B.R., 1990. Rice kernel moisture content variation at harvest. Transactions of the ASAE 33 (2), 541548. Leesawatwong, M., Jamjod, S., Kuo, J., Dell, B., Rerkasem, B., 2005. Nitrogen fertilizer increases seed protein and milling quality of rice. Cereal Chemistry 82, 588593. Lin, H.X., Liang, Z.W., Sasaki, T., Yano, M., 2003. Fine mapping and characterization of quantitative trait loci Hd4 and Hd5 controlling heading date in rice. Breeding Science 53, 5159. Liu, Z.H., Cheng, F.M., Zhang, G.P., 2005. Positional variations in phytic acid and protein content within a panicle of japonica rice. Journal of Cereal Science 41, 297303.
127
Lloyd, B.B., Cnossen, A.G., Siebenmorgen, T.J., 2001. Evaluation of two methods for separating head rice from brokens for head rice yield determination. Applied Engineering in Agriculture 17 (5), 643648. Lu, R., Siebenmorgen, T.J., 1995. Correlation of head rice yield to selected physical and mechanical properties of rice kernels. Transaction of the ASAE 38 (2), 889894. Luh, S., Luh, B., 1991. Properties of rice caryopsis. In: Rice Production, vol. 2. Van Nostrand Reinhold, New York. 394. McCauley, G.N., Way, M.O., 2002. Drain and harvest timing affects on rice grain drying and whole-milled grain. Field Crops Research 74, 163172. Pan, Z., Thompson, J.F., Amarantunga, K.S.P., Anderson, T., Zheng, X., 2005. Effect of cooling methods and milling procedures on the appraisal of rice milling quality. Transactions of the ASAE 48 (5), 18651871. Perez, C.M., Juliano, B.O., Liboon, S.P., Alcantara, J.M., Cassman, K.G., 1996. Effects of late nitrogen fertilizer application on head rice yield, protein content, and grain quality of rice. Cereal Chemistry 73, 556560. Reid, J.D., Siebenmorgen, T.J., Mauromoustakos, A., 1998. Factors affecting the slope of head rice yield vs. degree of milling. Cereal Chemistry 75, 738741. Samonte, S.O.P.B., Wilson, L.T., McClung, A.M., Tabien, R.E., 2006. Selection for stable high income grossing rice genotypes. In: Proceedings, 31st Rice Technical Working Group, p. 60. The Woodlands, Texas. Schluterman, D.A., Siebenmorgen, T.J., 2007. Relating rough rice moisture content reduction and tempering duration to head rice yield reduction. Transactions of the ASABE 50 (1), 137142. Seetanum, W., De Datta, S.K., 1973. Grain yield, milling quality and seed viability of rice as inuenced by time of nitrogen application and time of harvest. Agronomy Journal 65, 390394. Siebenmorgen, T.J., Jindal, V.K., 1986. Effects of moisture adsorption on the head rice yields of long-grain rough rice. Transactions of the ASAE 29, 17671771. Siebenmorgen, T.J., Counce, P.A., Lu, R., Kocher, M.R., 1992. Correlation of head rice yield with individual kernel moisture content distribution at harvest. Transactions of the ASAE 35, 18791884. Siebenmorgen, T.J., Bautista, R.C., Counce, P.A., 2007. Optimal harvest moisture contents for maximizing milling quality of long- and medium-grain rice cultivars. Applied Engineering in Agriculture 23 (4), 517527. Thompson, J.F., Mutters, R.G., 2006. Effect of weather and rice moisture at harvest on milling quality of California medium grain rice. Transactions of the ASABE 49 (2), 435440. Wang, F., Chen, S., Cheng, F., Liu, Y., Zhang, G., 2007. Differences in grain weight and quality within a rice (Oryza sativa L.) panicle as affected by panicle type and source-sink relation. Journal of Agronomy and Crop Science 193, 6373. Xiao, J., Li, L., Grandillo, S., Ahn, S.N., Yuan, L.P., Tanksley, S.D., McCouch, S.R., 1998. Identication of trait improving quantitative loci alleles from wild rice relatives, Oryza rupogon. Genetics 150, 899909. Xiong, L., Lui, K., Dai, X., Xu, C., Zhang, Q., 1999. Identication of genetics factors controlling domestication-related traits in rice using F2 population of a cross between Oryza sativa and O. rupogon. Theoretical and Applied Genetics 98, 243251. Zhang, Y., Luo, L., Xu, C., Zhang, Q., Xing, Y., 2006. Quantitative trait loci for panicle size, heading date and plant height co-segregating in trait performance derived near isogenic lines of rice (Oryza sativa). Theoretical and Applied Genetics 113, 361368. Zhuang, J.Y., Lin, H.X., Lu, J., Qian, H.R., Hittalmani, S., Huang, N., Zheng, K.L., 1997. Analysis of QTL x environment interaction for yield components and plant height. Theoretical and Applied Genetics 95, 799808.

Texture, processing and organoleptic properties of chickpea-fortied spaghetti with insights to the underlying mechanisms of traditional durum pasta quality
Jennifer Ann Wood*
NSW Department of Primary Industries, Tamworth Agricultural Institute, 4 Marsden Park Road, Calala, NSW 2340, Australia
Article history: Received 10 March 2008 Received in revised form 9 July 2008 Accepted 22 July 2008 Keywords: Cicer arietinum L Pasta quality Rheology Sensory
a b s t r a c t
Nutritionally enhanced spaghetti was prepared from durum semolina fortied with 030% desi chickpea besan our. This study examined the dough rheology, processing ease and quality attributes of the fortied spaghetti including protein, starch, texture (rmness, resilience and stickiness), colour, cooking loss, and organoleptic acceptability. Chickpea-fortied spaghetti was acceptable to consumers, had reasonable pasta quality, including lower cooking loss and less stickiness than the control spaghetti and retained rmness better than durum after refrigeration. This study suggests that chickpea-fortied spaghetti may be suited to uses such as fresh pasta, in soups, canning, and microwave re-heating. In addition, this study has added to the understanding of the underlying mechanisms of pasta quality. The main ndings were: (1) gluten content/composition appears to be more important than protein content for pasta rmness; (2) the proteinpolysaccharide matrix appears to be more important than the starch composition for cooking loss; (3) increased protein and amylose contents were associated with decreased pasta stickiness; (4) cooking loss and stickiness were not necessarily as strongly related as commonly believed. Further research into these theories is necessary to fully understand the underlying mechanisms of pasta quality. 2008 Elsevier Ltd. All rights reserved.
1. Introduction Nutritionists recommend pulses (grain legumes) such as chickpea (Cicer arietinum) in the diet as they have many nutritional benets. Chickpea seed has a high protein digestibility, contains high levels of complex carbohydrates (low glycaemic index), is rich in vitamins and minerals and is relatively free from anti-nutritional factors (Muzquiz and Wood, 2007; Wood and Grusak, 2007). Consumers are becoming increasingly health conscious and while many admit to knowing pulses are good for them, they are not sure how to use them in their diet. There is also a perception that cooking pulses is difcult and/or time consuming. Australian production of desi chickpea is currently around 200,000 tonnes per year (Knights et al., 2007), yet less than 0.5% of this is consumed by Australians. In comparison, pasta, produced from durum wheat (Triticum turgidum) our, is consumed worldwide. It is relatively non-perishable, inexpensive, easy to prepare, and readily accepted by all age groups. The nutritional benets, deep yelloworange cotyledon colour and lack of offensive beany avours compared to other pulses make chickpea a suitable candidate for incorporation with durum
* Tel.: 61 2 67631157; fax: 61 2 67631222. E-mail address: jenny.wood@dpi.nsw.gov.au 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.07.016
into pasta. In this way consumers could increase their consumption of pulses with little extra effort or thought. The inclusion of pulses in cereal based foods is known to increase the nutritive value by improving protein content and lysine availability (Bahnassey et al., 1986; Hernandez and Sotelo, 1984; Kurien et al., 1971; Reyes-Moreno et al., 2004; Siddique et al., 1996; Wood and Grusak, 2007). Several studies have examined various aspects of chickpea incorporation into pasta (Goni and Valentin-Gamazo, 2003; Sabanis et al., 2006) however the endproduct qualities of pasta produced from besan (dehulled desi chickpea our) has not been thoroughly investigated. Goni and Valentin-Gamazo (2003) showed that spaghetti containing 25% chickpea our had a signicantly lower glycaemic index (GI) than traditional durum spaghetti. Chickpea inclusion also increased the mineral and fat content without affecting the total starch content. Zhao et al. (2005) incorporated 5-30% of different pulse ours into spaghetti and found that rmness and colour intensity increased, while overall quality decreased. However, the chickpeas used in this study were kabuli types and the seed coats were not removed prior to grinding. Sabanis et al. (2006) investigated 5-50% inclusion of chickpea our in durum lasagne and found that the physical properties of the dough were improved; however, processing, handling and cooking characteristics deteriorated with the higher substitution levels.
J.A. Wood / Journal of Cereal Science 49 (2009) 128133
129
This study aimed to more comprehensively identify whether, besan, could be successfully incorporated into durum spaghetti to obtain a convenient health product with organoleptic properties acceptable to consumers. Furthermore, this research provides additional information to the existing literature on the effects of starch and protein on the quality of traditional durum pasta.
time, peak viscosity, breakdown, nal viscosity and setback were recorded. 2.5. Protein content and amino acid analysis Total protein content of the fortied products and controls were determined in duplicate with a LECO nitrogen analyser using AACC approved method 46-30 (AACC, 1995c). Protein determinations were based on nitrogen factors of 5.7 for wheat our and 6.25 for chickpea our. Nitrogen factors for spaghetti blends were adjusted on the basis of the relative proportion of proteins in the ingredients. Following extraction by acid hydrolysis, amino acid analysis of the 0%, 15% and 30% fortied spaghetti were determined by HPLC using AACC approved method 07-01 (AACC, 1995a). Tryptophan content was not determined. 2.6. Amylose content The amylose content of the ours and spaghetti were determined using the Megazyme amylose and amylopectin assay kit (Megazyme International, Ireland). 2.7. Sensory evaluation Sensory evaluation was performed on cooked spaghetti cooled in distilled water (5 min), drained, covered with plastic, and refrigerated overnight. Spaghetti fortied with %0, %15 and 30% chickpea our was randomly numbered and scored by 27 untrained volunteers the following day. Texture analysis was performed on this spaghetti as previously described. A hedonic scale was chosen for the sensory evaluation questionnaire. The volunteers were asked to give a ranking of the attributes (colour, rmness, avour, overall acceptability) for the three samples on a linear scale. The distance from the undesirable end of the scale was measured for each sample and subjected to ANOVA. 2.8. Statistical analyses All the parameters were analysed in duplicate. The data was statistically analysed by analysis of variance (ANOVA). Each parameter was tested for signicance (P < 0.05) between the fortied and control spaghetti samples. When signicant differences were found, the least signicant difference (LSD) test was used to determine the differences among means. A correlation coefcient between cooking loss and stickiness was determined using Microsoft Excel. 3. Results and discussion 3.1. Dough and processing properties Increasing chickpea fortication signicantly (P < 0.05) decreased water absorption and resulted in longer development times and less stable doughs (Table 1). Similar farinograph trends have been reported previously for wheat incorporating legume ours or their protein concentrates (Lorenz et al., 1979; Rasmay et al., 2000; Yanez-Farias et al., 1999). Many of these effects can be attributed to weakening of the gluten matrix due to the incorporation of chickpea our which contains no gluten. Chickpea proteins are comprised mainly of globulins (5360%) with lesser concentrations of albumins, prolamins and glutelins (Dhawan et al., 1991). Fortication made the dough particles increasingly sticky, causing them to aggregate during mixing. This made extrusion to produce the spaghetti increasingly difcult. For this reason, spaghetti production above 30% fortication was not undertaken. Chickpea our contains signicant levels of soluble non-starch
2. Experimental 2.1. Samples Unconditioned seed (50.0 g) of desi chickpea (cv. Amethyst) was dehulled in the Sheller component (attrition-style) of an SK Engineering mill (SK Engineering and Allied Works, Bahraich, India) to produce dhal (split cotyledons) according to Wood et al. (2008). The resulting dhal was subsequently milled through a 1.0 mm sieve in a Newport Scientic mill (Newport Scientic Pty Ltd, Narabeen, Australia) to obtain coarse, besan our. Commercial durum semolina was obtained from Goodman Fielders (Tamworth, Australia) and blended with 0, 10, 15, 20, 25 and 30% of the chickpea our by weight, in duplicate. No signicant difference was found between the moisture contents of the blended ours as determined by the AACC approved method 44-15A (AACC, 1995b). 2.2. Dough quality and pasta colour evaluation Dough properties were determined with a farinograph and constant dough weight using the AACC approved method 54-21 (AACC, 1995d). Pasta was produced under vacuum with standard 33.3% water addition and dried in a pasta oven at low temperature according to Wood et al. (2001). Pasta colour (L*, a*, b*) of multiple layers of parallel spaghetti strands were measured (mean of six readings) using a Minolta chroma meter CR-310 (Minolta Camera Co., Osaka, Japan) on both dry and cooked spaghetti. The change in colour due to cooking was determined by calculating
DE*
q DL 2 Da 2 Db 2
2.3. Cooked pasta quality evaluation The quality of the cooked spaghetti was evaluated according to Wood et al. (2001). Spaghetti strands (10 g, 7 cm lengths) were placed in 250 ml of vigorously boiling distilled water containing 5 ml of stock solution (0.125 g NaHCO3, 175 g NaCl in 500 ml distilled H2O). Optimum cooking time of the spaghetti was recorded when the core in the middle of the spaghetti strand was no longer visible when squashed between two perspex sheets. Texture analysis (rmness, resilience and stickiness) and cooking loss were performed on spaghetti cooked to optimum cooking time according to Wood et al. (2001). Cooking loss was calculated as described by Matsuo et al. (1992). 2.4. Rapid visco analyser (RVA) Spaghetti was milled into our through a 0.5 mm sieve in a Newport Scientic mill (Newport Scientic Pty, Ltd, Narabeen, Australia). Moisture content of the ours and spaghetti blends were determined by the AACC approved method 44-15A (AACC, 1995b). Pasting properties were evaluated in duplicate using 29 g of sample (13.8% solids, moisture corrected) in the Rapid Visco Analyser (Newport Scientic Pty, Ltd, Narabeen, Australia). The temperature was held at 25 C for 2 min, increased to 95 C over 5 min, held at 95 C for 3 min and decreased to 25 C over 5 min. Peak
130
Table 1 Effect of chickpea our fortication on spaghetti quality parameters Quality parameters Fortication (chickpea %) 0% Water absorption (%) DDT (min) B10 (BU) Protein (%) Glutamine/glutamicacid (g/100 g) Proline (g/100 g) Lysine(g/100 g) Cysteine (g/100 g) Methionine (g/100 g) Colour, L* (dry) Colour, a* (dry) Colour, b* (dry) Colour, L* (cooked) Colour, a* (cooked) Colour, b* (cooked) Colour change, DE Amylose (%) Peak time (min) Peak viscosity (cP) Breakdown (cP) Final viscosity (cP) Setback (cP) Firmness (g) Resilience (g) Stickiness (gs) Cooking loss (%) 57.6a 3.75d 50d 12.43e 4.81a 1.67a 0.22c 0.33a 0.21a 57.73a 3.56c 40.74a 78.37a 0.33e 32.52a 32.74a 23.01b 8.50a 2191c 922c 5976a 4707a 624.6a 62.28a 5.91a 5.15a 10% 57.6a 4.50c 60c 14.33d NA NA NA NA NA 57.25ab 6.25b 33.76b 75.85b 3.42d 27.03c 28.16b 22.52b 8.30a 2303b 957c 5802c 4456b 641.1ab 46.02ab 4.72bc 4.84ab 15% 57.2b 4.50c 70b 13.82d 4.48b 1.57b 0.36b 0.34a 0.21a 55.98abc 7.13ab 36.33ab 74.01c 4.66c 28.65bc 28.18b 23.63b 8.30a 2231c 959c 5891b 4619a 671.6a 51.81a 4.74bc 4.64ab 20% 56.3c 4.50c 70b 15.37c NA NA NA NA NA 55.20bc 7.61ab 36.72ab 73.10cd 5.21bc 29.95abc 27.07b 23.54b 7.73b 2246bc 965c 4532e 3250d 574.8bc 33.07bc 4.67bc 4.66ab 25% 55.8d 5.00b 75ab 16.99b NA NA NA NA NA 54.61c 8.00a 37.86ab 72.33d 6.10ab 30.27ab 27.21b 23.45b 7.70b 2518a 1089b 5345d 3916c 550.9cd 15.45c 4.81b 4.53b 30% 54.2e 5.75a 80a 17.42a 4.22c 1.51c 0.62a 0.33a 0.20a 53.96c 8.28a 38.74ab 71.26e 6.52a 30.61ab 27.19b 26.65a 7.70b 2529a 1180a 5407d 4058c 510.1d 13.93c 3.81c 4.50b 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.12 0.08 0.00 0.00 0.04 0.00 0.00 0.01 0.01 0.01 0.02 0.00 0.00 0.00 0.03 0.01 0.00 0.00 0.03 P
NA, not analysed. For each quality parameter, means within rows followed by different letters are signicantly different (P < 0.05) by least signicant difference (LSD) test.
polysaccharides (NSP), about ten times higher than bread wheat our (Naivikul and DAppolonia, 1979), and this may have contributed to the increased stickiness. Based on these observations and the lower farinograph water absorptions of the chickpeafortied doughs, it is possible that reducing the amount of water added may improve mixing and extrudability to some extent. 3.2. Dry pasta properties Protein content and amino acid composition generally increased (P < 0.05) with fortication. Protein content of control spaghetti was 12.4% compared to 17.4% for the 30% fortied spaghetti (Table 1). A similar trend was observed by Zhao et al. (2005). Chickpea blends showed increased levels of all amino acids except for cysteine and methionine (no signicant change) and glutamine/glutamic acid and proline which signicantly (P < 0.05) decreased with fortication (Table 1). Chickpea our is higher in protein but limiting in cysteine and methionine, hence no signicant change in their contents were observed in the fortied spaghetti. Glutamine/glutamic acid and proline are indicative of gluten-type proteins which explains their lower levels since chickpea our is gluten free. On the other hand, wheat is limiting in lysine. Lysine content signicantly (P < 0.05) increased by 64 and 182% in the 15 and 30% fortied spaghetti respectively (Fig. 1). The colour of all the dried spaghetti generally became signicantly (P < 0.05) less bright (lower L*), more red (higher a*) and less yellow (lower b*) as the percentage of chickpea our increased (Fig. 2). The 10% and 15% chickpea-fortied spaghetti were rated by the sensory panel as visually similar to the control, however the spaghetti fortied with 20% or more tended to display an undesirable brownish tint (supported by L*a*b* measurements, Table 1). This is similar to the ndings of Zhao et al. (2005). 3.3. Rapid visco analysis The chickpea besan our slurry had a much lower and atter RVA viscosity prole than the durum semolina slurry (Fig. 3). Wood
(2004) found similar comparisons in RVA proles between chickpea besan our (cv. Howzat) and a hard bread wheat our (cv. Drysdale). Chickpea has a lower starch content with a higher proportion of amylose and a greater non-starch polysaccharide (NSP) content and lipid content than durum and bread wheats (Fabriani and Lintas, 1988; Wood and Grusak, 2007). This probably explains much of the differences in the RVA proles of the besan and semolina slurries. The 10% and 15% fortied spaghetti displayed similar pasting proles to the control spaghetti. They generally had higher peak viscosities and lower nal viscosities than the control spaghetti (Table 1). The higher peak viscosities indicate larger water binding capacities of the starches in these blends. The 25% and 30% fortied spaghetti peaked earlier and higher than the control spaghetti with slightly lower nal viscosities and much lower setbacks. The slurries of these blends swelled faster and bound more water; however, less retrogradation occurred, resulting in a less viscous gel after cooking and cooling. Most of these observations can be explained by the higher NSP content of the fortied spaghetti. NSP has a high
0.7
17 Lysine Content Protein Content 16
Lysine Content (g / 100 g)
0.6 0.5 0.4 0.3 0.2 0.1 0.0
Protein Content (%)
15 14 13 12 11 10 0 15 30
% Chickpea in pasta
Fig. 1. Protein and lysine content of spaghetti increases with desi chickpea our fortication.
131
60 59 58
41 39
3.4. Cooked pasta properties The optimum cooking time was determined to be 10.5 minutes for all the spaghetti samples. After cooking, all the spaghetti samples were more bright (higher L*) and less yellow (lower b*) due to leaching and/or degradation of colour pigments such as carotenoids and xanthophyll. Bleaching was most pronounced in the control spaghetti, as calculated by DE* (Table 1). Despite this, the cooked chickpea-fortied spaghetti samples were generally less bright (L*), more red (a*) and less yellow (b*) than the control spaghetti (P > 0.05). However, there was no statistical difference between the 0%, 15% and 30% fortied spaghetti in panellist perceptions of cooked spaghetti colour. The 10% and 15% substituted spaghetti had similar rmness and resilience to the durum control spaghetti. Similarly, Zhao et al. (2005) found no signicant difference in the rmness of most of their chickpeaspaghetti blends compared to the durum control. However, our results showed that rmness and resilience decreased with fortication of 2030% (Table 1; Fig. 4). Increasing the protein content of durum spaghetti has been shown to increase rmness (Nobile et al., 2005; Sissons et al., 2005). In addition, lowering the amylose content has been shown to decrease rmness (Gianibelli et al., 2005). This suggests that the fortied spaghetti rmness should have increased as protein content and amylose content increased due to fortication. This appears to be the case up to a threshold of 15% fortication. However, chickpea our contains no gluten. As more chickpea our was added, the semolina gluten was effectively diluted (observed as a decrease in glutamine/glutamic acid contents; P < 0.05) leading to weakening of the gluten matrix and a decrease in spaghetti rmness. This result suggests that it may not be the protein content per se (nor the high amylose content), rather the gluten content and possibly gluten composition that may be more important in determining spaghetti rmness. This nding is consistent with the work of Sissons et al. (2005) who showed that gluten content increased spaghetti rmness with no consistent trend relating to glutenin/gliadin composition. The larger NSP content of the chickpea our may similarly contribute to weakening of the protein matrix. Legume starches generally contain more amylose than cereal starches. Zhao et al. (2005) found an increase in cooking loss (measured by solids loss) with fortication but this method is not comparable to amylose loss. The cooking loss test used in this study mainly measures the amount of amylose leached into the water during cooking (Matsuo et al., 1992). Sharma et al. (2002) and Gianibelli et al. (2005) found cooking loss to decrease in spaghetti with a low amylose content. Hence, it was initially presumed that the chickpea-fortied spaghetti would have increased cooking loss
b* (yellowness)
L* (brightness)
57 56 55 54 53 52 51 50 0 10 15 20 25 30 L b
37 35 33 31 29 27 25
% Chickpea in pasta (dry)

Fig. 2. Colour changes in dry spaghetti with increasing desi chickpea our fortication.
water absorbing capacity which has been shown to increase the viscosity of bread wheat slurries (Sasaki et al., 2000). On cooling, the NSP can impede the formation of a three dimensional starch network, preventing hydrogen bonds between amylose and amylopectin from forming during retrogradation, resulting in a less viscous gel (Kim and DAppolonia, 1977; Yoshimura et al., 1996). The 20% fortied spaghetti pasting curve did not track between the 1015% and 2530% spaghetti as expected (Fig. 3). Instead it peaked at a similar viscosity to the control and had a much lower nal viscosity. This indicates that the starch granules in the blend swelled similarly to the control, but again with less retrogradation and reduced viscosity after cooking and cooling. The sample was reanalysed twice more but all generated the same pasting curve. This unexpected result is difcult to explain. There may be some sort of synergistic process between the besan and semolina constituents occurring at 20% fortication.
120 8000 Semolina 80
Temp 'C
40 6000 0% Pasta 10% Pasta 30% Pasta
Viscosity cP
20% Pasta 4000
700 Chickpea Flour 650
70 60 50
Firmness (g)
600 550
Resilience
2000
40 30 500 450 Firmness (g) Resilience (g) 0 10 15 20 25 30 20 10 0
0 0 4 8 12 16 20
400
Time mins
Fig. 3. RVA prole of raw ours (durum semolina and desi chickpea course besan) and spaghetti with different levels of chickpea fortication.
% Chickpea in pasta
Fig. 4. General decrease in rmness and resilience of spaghetti with increasing desi chickpea our fortication.
132
due to the higher amylose contents. However, cooking loss was shown to decrease with fortication (Table 1; Fig. 5). This result suggests that amylose content is not the major factor in cooking loss. Sissons et al. (2005) found no relationship of cooking loss with protein content in durum spaghetti. However, our results showed decreased cooking loss with increased protein and NSP contents of fortied spaghetti. This suggests that the proteinpolysaccharide matrix (involving both starch and NSP) may be responsible for the retention of amylose during spaghetti cooking and not necessarily the starch composition per se, as suggested by Sissons et al. (2005). The chickpea-fortied spaghetti was also less sticky (P < 0.05) than the control (Fig. 5). Pasta surface stickiness is believed to be inuenced by both the surface structure of the spaghetti strand and starch exuded onto the strand surface during cooking (Cunin et al., 1995; Dexter et al., 1985; Perovic, 2000). Lowering the amylose content of spaghetti has been shown to increase stickiness (Gianibelli et al., 2005; Sharma et al., 2002). Furthermore, increasing the protein content of pasta has been associated with decreased stickiness (Nobile et al., 2005; Sissons et al., 2005). Hence, the reduced stickiness of the chickpea -fortied spaghetti is probably a result of both higher protein and higher amylose contents. However, whilst both the cooking loss and stickiness generally improved with chickpea fortication, they were only weakly correlated with one another (r 0.51). Since cooking loss, stickiness and rmness all decreased with the addition of besan chickpea our, it appears that gluten, per se, may have little effect on the retention of amylose and other carbohydrates during cooking. The proteinpolysaccharide matrix as a whole may be more likely to inuence spaghetti quality in terms of rmness, stickiness and cooking loss. Additional components of chickpea our that may aid the proteinpolysaccharide matrix in carbohydrate retention include monomeric proteins and starch-bound phospholipids. This clearly requires further investigation and the knowledge may be useful for genotype selection in durum wheat breeding programs. 3.5. Cooked and refrigerated pasta properties The sensory evaluation study gave an indication of consumer preference. Most panellists scored the 15 and 30% cooked fortied spaghetti to be equally acceptable as the durum control. No statistically signicant (P < 0.05) difference could be found for colour, avour or overall acceptability. However, the 30% blend was scored signicantly (P < 0.05) rmer than the 0% and 15% blends. This was unexpected, as the texture analyser results of freshly cooked spaghetti showed a decreasing rmness with higher chickpea fortication.
800.0 700.0 600.0
100% Durum pasta 20% Chickpea pasta
Firmness (g)
500.0 400.0 300.0 200.0 100.0 0.0 Normal Fridge
Fig. 6. Effect of refrigeration on the rmness of 100% durum spaghetti and spaghetti fortied with 20% desi chickpea our.
When re-assessed by the texture analyser, the panels ndings were conrmed. The fortied spaghetti were slightly rmer than the control after refrigeration (P < 0.05). Whilst all the spaghetti softened substantially after refrigeration, the effect was greater in the control durum spaghetti: a rmness decrease of 69% for the durum spaghetti compared to 66% for the fortied spaghetti (Fig. 6). Refrigeration differentially inuenced the chickpea-fortied spaghetti and durum control spaghetti to alter the rmness trend. Riva et al. (2000) found that the rate and degree of starch retrogradation was exaggerated at cold temperatures. This too could cause rmer spaghetti, if the chickpea-fortied spaghetti were more predisposed to retrogradation under cold temperatures then the durum control. However, the RVA results showed the fortied spaghetti to have less retrogradation than the durum control. No difference in spaghetti diameter between the durum and fortied spaghetti was detected. However, water absorption may have occurred by diffusion through the gelatinised system to cause an increase in weight with negligible volume change (Riva et al., 2000). If the chickpea blends were more susceptible to water absorption, spaghetti density would increase which may increase rmness. This is a possibility, as the RVA results showed the fortied spaghetti to have a larger water binding capacity during cooking and this may remain after cooling. This nding requires further investigation. 4. Conclusion Chickpea-fortied spaghetti was acceptable to consumers and provided an enhanced nutritional status via the amino acid prole. Lysine content increased by 64 and 182% in the 15 and 30% blends respectively whilst the total protein content and the content of most amino acids increased with fortication. Spaghetti processing and handling characteristics deteriorated as the level of fortication increased. Functional dough properties and spaghetti rmness were generally hindered by increasing amounts of chickpea our. However, spaghetti stickiness improved with increasing fortication and cooking loss was reduced. Chickpea-fortied spaghetti retained rmness much better than durum after refrigeration. A marketing advantage may exist if this desirable rmness is retained when blended pasta is subjected to canning, microwave re-heating and inclusion in soups. This rmness retention property may also be of interest to fresh pasta manufacturers. Chickpea-fortied pasta would make a cheap, attractive and convenient health food which is acceptable to consumers in western society. In addition, this
6.0 5.5
5.2 5.1
% Cooking Loss
Stickiness (gs)
5.0 4.9 4.8
5.0 4.5 4.0 3.5 0 10 15 20 25 30 Stickiness (gs) Cooking Loss (%)
4.7 4.6 4.5 4.4
% Chickpea in pasta
Fig. 5. General decrease in cooking loss and stickiness of spaghetti with increasing desi chickpea our fortication.
133
study has added to the understanding of the underlying mechanisms of pasta quality. The ndings were: (1) gluten content/ composition appears to be more important than protein content for pasta rmness; (2) the proteinpolysaccharide matrix appears to be more important than the starch composition for cooking loss; (3) supportive of previous ndings that increased protein and amylose contents are associated with decreased pasta stickiness; (4) cooking loss and stickiness are not necessarily as strongly related as commonly believed. Further research into these theories is necessary to fully understand the underlying mechanisms of pasta quality. Acknowledgements The author wishes to thank the Grains Research and Development Corporation (GRDC) for funding (DAN402), Goodman Fielders, Tamworth for providing semolina, L. Ayre (NSW Department of Primary Industries, Tamworth) for statistical analyses, and M. Sissons (NSW Department of Primary Industries, Tamworth) for use of the pasta extruder/dryer. References
American Association of Cereal Chemists (AACC), 1995a. Approved Methods of the AACC. Method 07-01, ninth ed. AACC, St Paul, MN. American Association of Cereal Chemists (AACC), 1995b. Approved Methods of the AACC. Method 44-15A, ninth ed. AACC, St Paul, MN. American Association of Cereal Chemists (AACC), 1995c. Approved Methods of the AACC. Method 46-30, ninth ed. AACC, St Paul, MN. American Association of Cereal Chemists (AACC), 1995d. Approved Methods of the AACC. Method 54-21, ninth ed. AACC, St Paul, MN. Bahnassey, Y., Khan, K., Harrold, R., 1986. Fortication of spaghetti with edible legumes. 1. Physicochemical, antinutritional, amino acid, and mineral composition. Cereal Chem 63, 210215. Cunin, C., Handschin, S., Walther, P., Escher, F., 1995. Structural changes of starch during cooking of durum wheat pasta. Lebensm. Wiss. Technol 28, 323328. Dexter, J.E., Matsuo, R.R., MacGregor, A.W., 1985. Relationship of instrumental assessment of spaghetti cooking quality to the type and the amount of material rinsed from cooked spaghetti. J. Cereal Sci 3, 3953. Dhawan, K., Malhotra, S., Dahiya, B., Singh, D., 1991. Seed protein fractions and amino acid composition in gram (Cicer arietinum). Plant Food Hum. Nutr 41, 225232. Fabriani, G., Lintas, C. (Eds.), 1988. Durum Wheat: Chemistry and Technology. AACC, St. Paul, MN. Gianibelli, M.C., Sissons, M.J., Batey, I.L., 2005. Effect of source and proportion of waxy starches on pasta cooking quality. Cereal Chem 82, 321327. Goni, I., Valentin-Gamazo, C., 2003. Chickpea our ingredient slows glycemic response to pasta in healthy volunteers. Food Chem 81, 511515. Hernandez, M., Sotelo, A., 1984. Nutritional evaluation of wheat our cookies supplemented with chickpea, cheese whey and amino acids. Nutr. Rep. Int 29, 845858. Kim, S.K., DAppolonia, B.L., 1977. Effect of pentosans on the retrogradation of wheat starch gels. Cereal Chem 54, 150160.
Knights, E.J., Aikgoz, N., Warkentin, T., Bejiga, G., Yadav, S.S., Sandhu, J.S., 2007. Area, production and distribution. In: Yadav, S.S., Redden, B., Chen, W., Sharma, B. (Eds.), Chickpea Breeding and Management. CAB International, Wallingford, UK, pp. 167178. Kurien, S., Narayanaswamy, D., Daniel, V.A., Swaminathan, M., Parpia, H.A., 1971. Supplementary value of pigeonpea (Cajanus cajan) and chickpea to poor diets based on kafr corn and wheat. Nutr. Rep. Int 4, 229236. Lorenz, K., Dilsaver, W., Wolt, M., 1979. Fababean our and protein concentrate in baked goods and in pasta products. Bakers Dig 53, 3945. Matsuo, R.R., Malcolmson, L.J., Edwards, N.M., Dexter, J.E., 1992. A colorimetric method for estimating spaghetti cooking losses. Cereal Chem 69, 2729. Muzquiz, M., Wood, J.A., 2007. Antinutritional factors. In: Yadav, S.S., Redden, B., Chen, W., Sharma, B. (Eds.), Chickpea Breeding and Management. CAB International, Wallingford, UK, pp. 143166. Naivikul, O., DAppolonia, B.L., 1979. Carbohydrates of legume ours compared with wheat our. 3. Nonstarchy polysaccharides. Cereal Chem 56, 4549. Nobile, M.A., Baiano, A., Conte, A., Mocci, G., 2005. Inuence of protein content on spaghetti cooking quality. J. Cereal Sci 41, 347356. Perovic, B., 2000. Role of starches in pasta cooking. Tecnica Molitoria 51, 726731. Rasmay, N.M.H., El-Shatanovi, G.A., Hassan, K.E.W., 2000. High-protein macaroni from legume ours and their protein concentrates. Ann. Agric. Sci 45, 555570. Reyes-Moreno, C., Cuevas-Rodriguez, E.O., Milan-Carrillo, J., CardenasValenzuela, O.G., Barron-Hoyos, J., 2004. Solid state fermentation process for producing chickpea (Cicer arietinum L.) tempeh our: physicochemical and nutritional characteristics of the product. J. Sci. Food Agric 84, 271278. Riva, M., Fessas, D., Schiraldi, A., 2000. Starch retrogradation in cooked pasta and rice. Cereal Chem 77, 433438. Sabanis, D., Makri, E., Doxastakis, G., 2006. Effect of durum our enrichment with chickpea our on the characteristics of dough and lasagne. J. Sci. Food Agric 86, 19381944. Sasaki, T., Yasui, T., Matsuki, J., 2000. Inuence of non-starch polysaccharides isolated from wheat our on the gelatinization and gelation of wheat starches. Food Hydrocolloids 14, 295303. Sharma, R., Sissons, M.J., Rathjen, A.J., Jenner, C.F., 2002. The null-4A allele at the waxy locus in durum wheat affects pasta cooking quality. J. Cereal Sci 35, 287297. Siddique, I.M., Awan, J.A., Salim-ur-RehmanGilani, A.H., Ali, M., 1996. Evaluation of composite our based on wheat (Triticum aestivum) and chickpea (Cicer arietinum) for chapati making. Sci. Int 8, 165168. Sissons, M.J., Egan, N.E., Gianibelli, M.C., 2005. New insights into the role of gluten on durum pasta quality using reconstitution method. Cereal Chem 82, 601608. Wood, J.A., 2004. RVA Viscosity of Chickpea Flours. Tech. J. Newport Sci 1, 1. Wood, J.A., Batey, I.L., Hare, R.A., Sissons, M.J., 2001. A comparison of Australian and imported spaghetti. Food Aust 53, 349354. Wood, J.A., Grusak, M.A., 2007. Nutritional Value of Chickpea. In: Yadav, S.S., Redden, B., Chen, W., Sharma, B. (Eds.), Chickpea Breeding and Management. CAB International, Wallingford, UK, pp. 101142. Wood, J.A., Knights, E.J., Harden, S., 2008. Milling performance in desi type chickpea (Cicer arietinum L.): effects of genotype, environment and seed size. J. Sci. Food Agric 88, 108115. Yanez-Farias, G.A., Bernal-Aguilar, V., Ramirez-Rodriguez, L., Barron-Hoyos, J.M., 1999. Fortication of some cereal foods with a chickpea protein concentrate. Food Sci. Tech. Int 5, 8993. Yoshimura, M., Takaya, T., Nishinari, K., 1996. Effects of konjac-glucomannan on the gelatinization and retrogradation of corn starch as determined by rheology and differential scanning calorimetry. J. Agric. Food Chem 44, 29702976. Zhao, Y.H., Manthey, F.A., Chang, S.K.C., Hou, H.J., Yuan, S.H., 2005. Quality characteristics of spaghetti as affected by green and yellow pea, lentil, and chickpea ours. J. Food Sci 70, S371S376.

Impact of re-grinding on hydration properties and surface composition of wheat our

M. Mohamad Saad a, C. Gaiani a, *, J. Scher a, B. Cuq b, J.J. Ehrhardt c, S. Desobry a
a
LSGA, Laboratoire de Sciences et Genie Alimentaires, Nancy University, 2 avenue de la Foret de Haye, B.P. 172, 54505 Vandoeuvre Les Nancy Cedex, France Montpellier SupAgro, INRA, UMR 1208 Unit for Emerging Technology and Polymer Engineering, 2, place Viala, 34060 Montpellier Cedex 1, France c ` LCPME, Laboratoire de Chimie Physique et de Microbiologie pour lEnvironnement, Nancy Universite, CNRS, 405, rue de Vandoeuvre, 54600 Villers Les Nancy, France
b
Article history: Received 30 November 2007 Received in revised form 28 July 2008 Accepted 5 August 2008 Keywords: DVS Sorption isotherm Wheat our XPS
a b s t r a c t
The combination of two analytical methodologies (water vapor sorption isotherm by using the DVS and chemical surface composition by using the XPS) has been used to enhance the understanding of the impact of re-grinding on the wheat our hydration mechanism. A controlled atmosphere microbalance was used to construct water sorption isotherms at 25 C of different samples of wheat ours obtained by successive re-grinding of native wheat our. Experimental water adsorption isotherms were modeled using different complementary models, based on two-parameter (BET), three-parameter (GAB), and four-parameter (TSS) models. A slight increase in water sorption capacity of wheat our due to the re-grinding process was observed. The most affected parameters of the sorption isotherm models were C (the energy constant) and Xm (the monolayer water content capacity). The X-ray photoelectron spectroscopy (XPS) analysis showed changes in chemical bonds on wheat particle surfaces due to re-grinding process and particularly a signicant increase in hydrophilic and decrease in hydrophobic bonds. 2008 Elsevier Ltd. All rights reserved.
1. Introduction One of the major control variables in food preservation technology is water activity (aw). This term indicating the quality of water content in food, describes the degree of boundness of water and thus its availability to participate in physical, chemical, and microbiological reactions (Brunauer et al., 1938; Van den Berg and Bruin, 1981). The moisture sorption isotherm is dened as the relationship between the total moisture content and water activity of the food at a constant temperature and under equilibrium conditions. Moisture sorption isotherms are useful, not only in showing at which moisture content certain desirable or undesirable levels of aw is achieved, but also in indicating what signicance small changes in moisture content will have in terms of changes in aw. It is therefore a useful guide to the storage life of foods held at moderate temperatures and preserved only by reduced aw (Abdullah et al., 2000). The mathematical description of the sorption phenomena in foods is of interest. More than 200 equilibrium moisture content equilibrium relative humidity relationships have been reported in
Abbreviations: BET, BrunauerEmmettTeller; DVS, dynamic vapor sorption; GAB, GuggenheimAndersende Boer; RH, relative humidity; TSS, third stage sorption; XPS, X-ray photoelectron spectroscopy. * Corresponding author. Tel.: 33 3 83 59 58 78; fax: 33 3 83 59 58 04. E-mail address: claire.gaiani@ensaia.inpl-nancy.fr (C. Gaiani). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.08.001
the literature (McMinn et al., 2004). Both the International Union of Pure and Applied Chemistry (IUPAC) and the Commission on Colloid and Surface Chemistry (report of 1985) recommend the usefulness of the two-parameter model BET (Brunauer, Emmett, and Teller) because of its simplicity of application. This model is based on a multilayer sorption and is used in the low water activity range (0.05 < aw < 0.40). A three-parameter model (GAB model) was recommended by the European Project Group COST 90 (Wolf et al., 1984) and has been successfully applied to various foods (Van den Berg, 1985). The GAB (Guggenheim, Andersen, and de Boer) equation represents a model based on multi-layers and condensation and covers a wider range of water activity (0.05 < aw < 0.8) (Chirife et al., 1992; Timmermann and Chirife, 1991; Van den Berg and Bruin, 1981). Timmermann (1989) and Timmermann and Chirife (1991) developed a four-parameter equation, named the third stage sorption isotherm (TSS) to extend the GAB isotherm model to water activity ranges approaching unity. It is based on the premise that after a certain number of moisture layers exist, the moisture behaves as liquid water which has a dilution effect. The TSS model has thus the ability to predict innite moisture adsorption (Bronlund and Paterson, 2004) and to give experimental data for many food systems a better t when compared to the GAB equation. At the experimental level, the water adsorption capacity of food products has recently been evaluated using controlled atmosphere microbalances and dynamic automated sorption methods (Bell and Labuza, 2000; Johnson and Brennan, 2000). The
M.M. Saad et al. / Journal of Cereal Science 49 (2009) 134140
135
small size of samples and the dynamic airow around the samples enable generation of a complete isotherm in less than a week (Bell and Labuza, 2000). Several studies using controlled atmosphere microbalances have already been reported (Ketal et al., 2004; Levoguer and Willians, 1997) The surface composition and particle properties can be considered as critical parameters of powders as they are supposed to play an important role within the powders end use processes and should certainly not be neglected (Faldt, 1995; Kim et al., 2002). Understanding the mechanism of the powder surface formation in terms of the compositional aspect will be highly useful in powder quality improvement and new product development (Kim et al., 2002). A classical way to characterise the powders surface composition is the use of X-ray photoelectron spectroscopy, also referred in the literature to electron spectroscopy for chemical analysis (ESCA) (Mistry et al., 1992). XPS is a well established technique for identifying elements and determining differences in surface chemistry of particle surfaces (Sionkowska et al., 2006). XPS has been commonly applied to spray-dried dairy powders (skim milk powder, whole milk powder, cream powder and whey protein concentrate) in order to investigate the relationship between milk powder processing, particle surface structure, and wetting properties to make advances in the understanding of the rehydration process or storage effects (Gaiani et al., 2006, 2007; Kim et al., 2002). The main objective of the current work was to study the impact of the re-grinding process on hydration properties of wheat our. This study is primarily concerned with determining the experimental sorption data of different re-ground wheat our samples coming from the same batch of wheat our and describing them by using the two-, three-, and four-parameter isotherm models. Water adsorption measurements by using the dynamic vapor sorption have been conducted in combination with measurement of changes in surface bonds on the surface of wheat our particles using the Xray photoelectron spectroscopy. 2. Materials and methods 2.1. Wheat our sample preparation and conservation The different samples were prepared from the selected commercial wheat our (Grands Moulins de Paris, Ivry Sur Seine, France) intended for bread making (extraction rate 75%). Initial water content of the native wheat our (FNative) was 13.7%. The different our samples (F1, F2, and F3) were obtained by successive re-grinding (R1, R2, and R3) of the native our to reduce particle size using a laboratory grinder, under ambient relative humidity conditions (ZM 200, Retsch, France). For the rst re-grinding process R1, a 50 g sample of native wheat our was placed inside the laboratory grinder. The resulting wheat our was named F1. The our F1 was twice re-ground under the same conditions R2 and R3, and then the re-ground ours F2 and F3 were produced. During our study, all samples were stored at 18 C in hermetically sealed cans, in order to preserve their physicochemical characteristics (Da Costa-Correia, 1997; Kusunose et al., 2002). Before use, samples were defrosted for 12 h at 20 C. 2.2. Wheat our characterization 2.2.1. Chemical analyses Water and damaged starch contents of our samples were estimated according to American Association of Cereal Chemists (AACC) methods 44-15A and 76-30A, respectively. The total nitrogen content (TN) was determined by the Kjeldahl method and protein content was calculated according to TN 5.7 AACC (2000) Method (No: 46-10). Fat content was determined by acid hydrolysis
with HCl followed by extraction of hydrolysed lipids with ethers according to AACC Method (No: 30-10). All results represent the average of three tests. 2.2.2. Physical properties The particle size distributions were measured by static light scattering (Mastersizer S, Malvern Instruments Ltd, Malvern, UK) with a 5 mW HeNe laser operating at a wavelength of 632.8 nm with a 300F lens. Five grams of wheat sample were taken and introduced in 100 mL ethanol to reach a correct obscuration (Berton et al., 2002). The criterion selected was d50 which means that 50% of the particles have a diameter lower than this criterion (i.e. midpoint of volume cumulative distribution). Results are the average of triplicate experiments carried out on different days. 2.3. Sorption isotherms Water sorption isotherms were determined gravimetrically using DVS technique (Surface Measurement Systems, London, UK). The DVS apparatus monitors the moisture sorption capacities of wheat our as function of relative humidity (RH). The changes in sample weight over time at 25 C and at any desired RH (between 0% and 98%) were recorded. About 1520 mg of sample were loaded onto the quartz sample pan. The program was initially set to control the humidity at 0% for 12 h (drying phase). This step allowed the sample water activity to decrease to zero and internally equilibrate. The sample was then subjected to successive steps of 10% RH increase, up to 98%. For each step, mass changes (m) and the rate of mass changes (dm/dt) were plotted against time. The equilibrium was considered to be reached when changes in mass with time (dm/dt) were lower than 0.002%/min (i.e. 2 g water/100 g db/day). All experiments were run at 25 C and 24 tests were carried out for each sample. The accuracy of the system was 1.0% RH and 0.2 C, respectively. 2.4. XPS analysis The XPS analyses were carried out with a Kratos Axis Ultra (Kratos Analytical, Manchester, UK) spectrometer using a monochromatic Al Ka source. The delay-line detector allows a high count rate and the power applied to the X-ray anode was reduced to 90 W in order to avoid the X-ray induced degradation of the sample. The instrument work function was calibrated to give a binding energy (BE) of 83.96 eV for the Au 4f7/2 line for metallic gold and the spectrometer dispersion was adjusted to give a binding energy of 932.62 eV for Cu 2p3/2 line for metallic copper. The wheat our samples were attached to the sample holder using a double sided conductive tape and then evacuated overnight prior to analyses. All spectra were recorded at a 90 take-off angle, the analysed area being currently about 700 300 mm. Survey spectra were recorded with 1.0 eV steps and 160 eV analyser pass energy and the high resolution regions with 0.05 eV steps and 20 eV pass energy. In both cases the hybrid lens mode was used. During the data acquisition the Kratos charge neutralizer system was used on all specimens with the following settings: lament current 1.6 A, charge balance 2.4 V, lament bias 1.0 V and magnetic lens trim coil 0.375 A. As overcompensation is always observed, the C1s line for adventitious carbon and CC carbon was set to 284.60 eV and therefore used as an internal energy reference. With these parameters we could obtain a C1s signal with sharp, symmetric components with a FWHM of 1.2 eV. Spectra were analysed using the Vision software from Kratos (Vision 2.2.2). A Shirley baseline was selected to subtract the background and GaussianLorentzian (7030%) shapes were used for spectral decomposition. Quantication was performed using the photoemission cross-sections and the transmission coefcients given in the Vision package.
136 Table 1 Wheat our samples chemical composition Wheat our Composition (g/100 g our) Water FNative F1 F2 F3 13.7 0.1 13.6 0.2 13.4 0.2 13.4 0.3 Protein 12.5 0.1 12.4 0.1 12.4 0.2 12.4 0.3
M.M. Saad et al. / Journal of Cereal Science 49 (2009) 134140 Table 2 Wheat our samples particle size distribution Wheat our Damaged starch 4.7 0.4 8.5 0.0 9.1 0.4 12.6 0.8 Lipids 1.5 0.2 1.4 0.1 1.6 0.1 1.2 0.3 FNative F1 F2 F3 d10 (mm) 15.1 0.8 11.4 2.6 10.0 2.6 4.3 0.9 d50 (mm) 67.4 0.6 49.7 4.6 40.4 4.6 17.2 3.8 d90 (mm) 147.8 1.4 122.2 3.7 108.4 3.7 44.5 3.8
2.5. Data analysis The water vapor adsorption isotherms were described by using three models: the BET (BrunauerEmmettTeller) model (Eq. (1)), GAB (GuggenheimAndersende Boer) model (Eq. (2)), and TSS (third stage sorption) model (Eq. (3ac)).
H0 aw h1
Haw 1 1 KTSS aw hTSS 1 hTSS aw Haw 1 aw
(3c)
X Xm
CBET aw 1 aw 1 aw CBET aw
(1)
X Xm
CGAB KGAB aw 1 KGAB aw 1 KGAB aw CGAB KGAB aw
(2)
X Xm
Haw H0 aw CTSS KTSS aw 1 KTSS aw 1 CTSS Haw 1KTSS aw
(3a)
where Xm is the monolayer moisture content (% db), X is the equilibrium water content (% db), CBET, CGAB, and CTSS are characteristic energy constants, KBET, KGAB, KTSS are the characteristic constants correcting the properties of the multilayer molecules with respect to the bulk liquid, and hTSS is the third stage sorption isotherm constant. The model parameters were directly determined from experimental water vapor sorption isotherms. The parameters of the BET model (Xm, C), GAB model (Xm,C, and K), and TSS model (Xm, C, K, and hTSS) were identied, respectively, with Eqs. (1)(3). The parameters were calculated by an optimization procedure according to the GaussNewton algorithm using the software Excel 2007 (Microsoft). The minimized objective function was the sum of the absolute difference between experimental and predicted points. 2.6. Statistical analyses
1 KTSS KTSS aw hTSS Haw h1 1 aw KTSS
(3b)
All statistical analysis was carried out by using Microsoft Excel 2007 (Microsoft Corporation, USA).
Fig. 1. Sorption isotherm prole obtained for the FNative, F1, F2 and F3 wheat our samples estimated at 25 C from 0% to 98% RH and modeled up with the TSS model.
M.M. Saad et al. / Journal of Cereal Science 49 (2009) 134140 Table 3 Parameters obtained from the tted curves with BET, GAB, and TSS models for FNative, F1, F2 and F3 wheat our samples Model BET Flour FNative F1 F2 F3 FNative F1 F2 F3 FNative F1 F2 F3 Xm 6.41 0.69 6.28 0.46 6.36 0.03 6.47 0.08 8.50 0.75 8.36 0.94 8.54 0.02 8.64 0.04 8.84 1.14 8.22 1.21 9.32 3.19 9.09 0.64 C 18.9 10.1 18.2 1.2 17.4 1.0 21.1 2.2 12.9 4.9 13.4 1.5 12.8 0.5 13.6 0.3 11.4 3.0 14.5 3.0 10.9 0.1 11.5 1.6 K 0.686 0.015 0.701 0.046 0.678 0.014 0.609 0.008 0.665 0.037 0.710 0.067 0.634 0.036 0.678 0.042 hTSS 14.3 1.9 17.1 5.8 17.5 1.1 14.4 0.5 R2 0.992 0.990 0.994 0.994 0.997 0.997 0.998 0.997 0.998 0.988 0.997 0.998
137
Abs. difference 0.266 0.296 0.249 0.263 0.159 0.184 0.138 0.190 0.234 0.178 0.290 0.224
GAB
TSS
Xm: monolayer moisture content g of water/100 g of dry base; C: constant; K: constant; hTSS: TSS model constant; Abs. difference: the average value of absolute difference between experimental data and calculated value of water content.
3. Results and discussion 3.1. Wheat our characterization The chemical composition of the selected samples of wheat our is presented in Table 1. As reported by Berton et al. (2002) and Wang and Flores (2000), the re-grinding process produced a signicant increase in damaged starch content (from 5% to 13%), while the protein and lipid contents remain stable (about 12.5% and 1.5%, respectively). The water content slightly decreases from 13.7% to 13.2% with re-grinding and this could be due to the slight heat effects during the re-grinding process (10 C). As expected, the size distribution of our samples decreases as re-grinding progresses (Table 2). The d50 of the samples is around 67 mm for native our (FNative) and decreases to 50, 40, and 17 mm, respectively, for F1, F2 and F3 ours. The decrease in particle size with re-grinding is due to the successive breaking of our particles inside the grinder. The breaks in particles occurring during re-grinding have been considered as responsible for starch granule rupture and thus induce increase in damaged starch content (Hoseney, 1994). According to Dubois (1949), starch granule exhibits elastic properties that lead to different types of damage such as cracks and breaks during grinding, which play an essential role in increasing the surface subjected to water vapor. 3.2. Water vapor adsorption isotherms The adsorption isotherm proles at 25 C, from 0% to 98% RH are presented in Fig. 1 for the selected wheat our samples. As expected, these isotherms demonstrate an increase in equilibrium moisture content with increasing water activity. As has been shown by Roman-Gutierrez et al. (2002), this behavior is obviously a sigmoidal shaped curve reecting a Type II isotherm according to the Brunauer classication (Brunauer et al., 1938). This form can be described in terms of a three-step moisture adsorption process. As previously observed by other researchers (Quirijns et al., 2005), sorption isotherms involve, during the rst step, polar groups of high binding energy to hydrophilic components (starches, proteins, and pentosans) being saturated with water molecules (i.e. the monolayer coverage). During the second stage, additional water molecules are bound onto the monolayer (i.e. multilayer coverage) and water clusters begin to form. During the third step, accumulation of water in intermolecular free spaces occurs and results in partial swelling that in turn may expose additional hydrophilic binding sites. It can be noticed that very slight differences seem to be observed between the experimental adsorption isotherms for the native wheat ours and the three re-ground our samples.
3.3. Water vapor sorption mathematical modeling 3.3.1. BrunauerEmmettTeller Model The experimental sorption data of wheat our samples were rst tted with the two parameter BET model (Eq. (1)), between 0% and 40% RH (Table 3). There is a good agreement between experimental data and predicted values (average R2 0.992 and average absolute difference between experimental and calculated data 0.27%). The calculated Xm (monolayer moisture content) and CBET values are presented in Table 3 for native and re-ground our samples. No differences have been found between the calculated values of BET model parameters for the four different products. We can report almost the same values of Xm (6.286.47 g/100 g dry bases) for the native wheat our and the re-ground our samples. Almost
Fig. 2. XPS spectra obtained for the FNative wheat our (survey scan).
138
the same values for monolayer water contents (Xm) have already been reported in the literature for wheat ours (Roman-Gutierrez et al., 2002). We observe a slight increase in the CBET values with re-grinding, from CBET 18.9 for the native our to CBET 21.1 for the sample F3. That is to say, when using the BET model, the energy constant (CBET) is found to be slightly increased as re-grinding rate increases. This behavior may be manifested by the liberation of hydrophilic sites during the re-grinding process, whereas different types of starch damage as cracks and breaks take place. 3.3.2. GuggenheimAndersende Boer Model The experimental sorption data of wheat our samples have also been tted with the three-parameter GAB model (Eq. (2)), between 0% and 80% RH (Table 3). As expected, there is good agreement between experimental data and predicted values (average R2 0.997 and average absolute difference between experimental and calculated data 0.17%). The calculated GAB model parameter Xm (monolayer moisture content), KGAB, and CGAB values are presented in Table 3 for the native wheat our and the three re-ground our samples. No signicant differences have been found between the calculated values of GAB model parameters for the four products. We can report almost the same values of Xm (8.368.64 g/100 g dry bases) for the native wheat our and the re-ground samples, nonetheless the highest values were observed for the most re-ground our (sample F3). It can be noticed that the
Xm values determined using the GAB model (8.368.64 g/100 g dry bases) are slightly higher than those calculated using the BET model (6.286.47 g/100 g dry bases). Timmermann (2003) also stated that the value Xm given by the BET model is always smaller than the monolayer value corresponding to the GAB model. No differences were observed among CGAB values (12.813.7 g/ 100 g dry bases) for the native wheat our and the re-ground our samples. Still, the highest values were observed for the most re-ground our (sample F3). The KGAB values remain constant (between 0.678 and 0.701) for all our samples. The estimated value conrms observations, carried out on many starchy foods (Chirife et al., 1992), which reveals that KGAB values are about 0.74. Probably, this stability in KGAB values during the re-grinding process may be attributable to the steady chemical composition of all studied our samples. Similar values of Xm, CGAB, and KGAB have already been reported in the literature for wheat our (Chirife et al., 1992; Roman-Gutierrez et al., 2002). 3.3.3. Third stage sorption model The experimental sorption data of wheat our samples have then been tted with the four-parameter TSS model (Eq. (3ac)) between 0% and 98% RH as illustrated in Fig. 1 and Table 3. There is good agreement between experimental data and predicted values (average R2 0.995 and average absolute difference between experimental and calculated data 0.23%). The calculated TSS
Fig. 3. Example of XPS narrow spectra of O1s obtained for FNative, F1, F2 and F3 wheat our samples, respectively.
139
model parameters Xm (monolayer moisture content), CTSS, KTSS, and hTSS values are presented in Table 3 for the native wheat our and the three re-ground our samples. The calculated values of the TSS model (Xm, CTSS, KTSS, and hTSS) appear to stay constant for all the selected samples, with Xm values ranging between 8.84 and 9.32 g/100 g dry bases, CTSS values ranging between 10.9 and 14.5, KTSS ranging between 0.634 and 0.710, and with hTSS values ranging between 14.3 and 17.5. Therefore, it may be stated that no signicant change was observed between the native wheat our and the three re-ground our samples, when describing the whole adsorption isotherms between 0% and 98% relative humidity with the TSS model. The calculated values of the hTSS parameter appear to be low in comparison with usual values that have been found (Timmermann, 1989). Bronlund and Paterson (2004) have found that hTSS value is around 30. Because of wheat our insolubility in water (nite sorption at RH 100%), the TSS model takes into consideration the relative insolubility behavior of wheat components by lowering the hTSS value. This could explain why low values of h parameter have been obtained. 3.4. Wheat our chemical surface analyses using X-ray photoelectron spectroscopy The survey scan of the native wheat our sample is represented as an example with the identication of the O1s, N1s, C1s, S2p and P2p peaks as illustrated in Fig. 2. The detection of S2p and P2p elements was possible thanks to the high sensitivity of the XPS equipment (Gaiani et al., 2006). Nevertheless, the concentration of these two elements was found below 1% for all the samples. Almost the same XPS curve shapes have been previously observed for milk powders (Gaiani et al., 2006). According to a model for biochemical compounds, the carbon (C), oxygen (O) and azotes (N) peaks were decomposed (Gerin et al., 1995). The C1s peak was decomposed into four peaks corresponding to the C(C, H), C(O, N), C]O and O C]O functions. The O1s peak was decomposed in three peaks attributed to the O]C, OC and H2O functions. The N1s peak was decomposed into CNH and CNH3 functions. As a typical example, decomposition of the oxygen peak for the native wheat our and the three re-ground our samples are presented in Fig. 3. Similar decomposition curves have been constructed for carbon and nitrogen (data not shown). From these decompositions, the data were analysed in terms of rate of surface bonds (Table 4). From the C1s peak, it appears that the re-grinding process (from FNative, F1, F2, to F3 our) induces a slight decrease of
Table 4 Elemental surface composition (bold) and rate of surface bond for FNative, F1, F2 and F3 wheat our samples Element Bond Wheat our sample FNative %C % % % % %O % OC % O]C % H2O %N % CNH % CNH3 %S %P CH, CC CO, CN C]O OC]O 77.0 61.0 28.2 6.2 4.6 18.6 82.0 13.7 4.3 3.9 93.2 6.8 0.3 0.2 F1 77.4 60.1 27.9 8.6 3.4 18.0 77.5 14.1 8.4 4.1 92.7 7.3 0.3 0.2 F2 76.7 59.1 29.3 8.0 3.6 18.9 76.9 15.2 7.9 4.0 96.5 3.5 0.2 0.2 F3 74.7 57.2 31.0 8.7 3.1 19.3 75.2 18.9 5.9 5.5 93.6 6.4 0.3 0.2
Fig. 4. Changes in surface chemical bonds percentage as re-grinding rate increases for FNative, F1, F2 and F3 wheat our samples, where white columns represent C(H,C) and black columns represent C(ON), C]O, OC]O.
the C(C, H) bond (from 61, 60, 59 to 57%, respectively, for FNative, F1, F2, to F3 samples). Concurrently the total of CO functions (C(O, N), C]O and OC]O) increased. From the O1s peaks, the O]C functions increased signicantly whereas the OC functions decreased (from 82.0, 77.5, 76.9 to 75.2, respectively, for FNative, F1, F2, to F3 samples). These results clearly reect a decrease in the number of hydrophobic bonds in contrast to hydrophilic bonds which increase in its turn (Fig. 4). This behavior may be due to physical changes induced by re-grinding process resulting in starch granule rupture. 4. Conclusion Whatever the model applied (BET, GAB, or TSS), wheat our shows a slight trend to adsorb more water vapor after the regrinding process. This bonded water was adsorbed more strongly. One could assume that starch damages occurring during the regrinding process could affect the surface and modify interactions of particles with water vapor, whereas the hydrophilic bonds increase. Hence, the TSS (third stage sorption) model can be used coupled with the DVS technique to describe sorption isotherms of wheat our throughout the entire range of water activity. Techniques such as DVS can be used coupled with XPS to obtain valuable information related to total and partial contribution of chemical components of wheat ours in their hydration properties. Therefore, further studies including surface composition evolution during milling coupled with sorption isotherms of pure wheat our components (starch, protein,.) would improve understanding contributions of wheat our components with water vapor adsorption. Acknowledgements We are grateful to J. Lambert, LCPME engineer CNRS (Nancy), for providing XPS analysis and technical assistance. The authors would like also to acknowledge the support of Syrian Ministry of Higher Education for nancial support. References
American Association of Cereal Chemists, 2000. Ofcial Methods of the AACC. Method 44-15A approved October 1975, revised October 1981; revised November 1999; Method 76-30A, approved May 1969, revised October 1982, and October 1984; revised November 1999; Method 46-10, approved April 1961, revised November 1999; Method 30-10, approved April 1961, revised November 1999, 10th ed. The Association, St. Paul, MN, USA. Abdullah, N., Nawawi, A., Othman, I., 2000. Fungal spoilage of starch-based foods in relation to its water activity (aw). Journal of Stored Products Research 36, 4754.
140
M.M. Saad et al. / Journal of Cereal Science 49 (2009) 134140 structure and wetting properties. Colloids and Surfaces B: Biointerfaces 26 (3), 197212. Kusunose, C., Noguchi, S., Yamagishi, T., Seguichi, M., 2002. Binding starch to tailing fraction by proteins in stored wheat our. Food Hydrocolloids 1, 7377. Levoguer, C.L., Willians, D.R., 1997. Moisture sorption properties of food products and packaging materials studied by dynamic vapour sorption. Food Technology Europe 6, 2830. McMinn, A.M., Al-Muhtaseb, A., Ronald, A., Magi, T., 2004. Assessment of two- and three-parameter Lewicki models for description of sorption phenomena of starch materials. Journal of the Science of Food and Agriculture 84, 16951700. Mistry, V.V., Hassan, H.N., Robinson, D.J., 1992. Effect of lactose and protein on the microstructure of dried milk. Food Structure 11, 7382. Quirijns, E.J., Boxtel, A.J.B.V., Straten, W., 2005. Sorption isotherms, GAB parameters and isosteric heat of sorption. Journal of the Science of Food and Agriculture 85, 18051814. Roman-Gutierrez, A., Guilbert, S., Cuq, B., 2002. Distribution of water between wheat our components: a dynamic water vapor adsorption study. Journal of Cereal Science 36, 347355. Sionkowska, A., Wisniewski, M., Kaczmarek, H., Skopinska, J., Chevallier, P., Mantovani, D., Lazare, S., Tokarev, V., 2006. The inuence of UV irradiation on surface composition of collagen/PVP blended lms. Applied Surface Science 253, 19701977. Timmermann, E.O., 1989. A BET-like three sorption stage isotherm. Journal of Chemical Society 85, 16311645. Timmermann, E.O., Chirife, J., 1991. The physical state of water sorbed at high activities in starch in terms of the GAB sorption equation. Journal of Food Engineering 13, 171179. Timmermann, E.O., 2003. Multilayer sorption parameters: BET or GAB values? Colloids and Surfaces A: Physicochemical and Engineering Aspects 220, 235260. Van den Berg, C., Bruin, S., 1981. Water activity and its estimation in food systems. In: Rockland, L.B., Stewart, F. (Eds.), Water Activity: Inuence on Food Quality. Academic Press, New York, pp. 147177. Van den Berg, C., 1985. Development of B.E.T.-like models for sorption of water on foods, theory and relevance. In: Simatos, D., Multon, J.L. (Eds.), Properties of Water in Foods. Martinus Nijhoff, Dordrecht, p. 119. Wang, L., Flores, R.A., 2000. Effects of our particle size on the textural properties of our tortillas. Journal of Cereal Science 31, 263272. Wolf, W., Spiess, W.E.L., Jung, G., Weisser, H., Bizot, H., Duckworth, R.B., 1984. The water vapor sorption isotherms of microcrystalline cellulose (MCC) and of puried potato starch: results of a collaborative study. Journal of Food Engineering 3, 5172.
Bell, L.N., Labuza, T.P., 2000. Moisture Sorption: Practical Aspects of Isotherm Measurement and Use. American Association of Cereal Chemists, Inc., St Paul, MN. Berton, B., Scher, J., Villieras, F., Hardy, J., 2002. Measurement of hydration capacity of wheat our: inuence of composition and physical characteristics. Powder Technology 128, 326331. Bronlund, J., Paterson, T., 2004. Moisture sorption isotherms for crystalline, amorphous and predominantly crystalline lactose powders. International Dairy Journal 14, 247254. Brunauer, S., Emmett, P.H., Teller, E., 1938. Adsorption of gases in multimolecular layers. Journal of American Chemical Society 60, 309319. Chirife, J., Timmermann, E.O., Iglesias, H.A., Boquet, R., 1992. Some features of the parameter k of the GAB equation as applied to sorption isotherms of selected food materials. Journal of Food Engineering 15, 7582. ` Da Costa-Correia, J.M., 1997. Contribution a letude des proprietes rheologiques des farines de ble: Inuence de la composition, de lhydratation de la distribution ` granulometrique des particules. Ph.D. thesis, Institue Nationale Polytechnique de Lorraine. Dubois, M., 1949. Incidences du reglage de la mouture sur les proprietes des farines. Bulletin ENSMIC 113, 170187. Faldt, P., 1995. Surface Composition of Spray-dried Emulsions. Ph.D. thesis, Lund University. Gaiani, C., Ehrhardt, J.J., Scher, J., Hardy, J., Desobry, S., Banon, S., 2006. Surface composition of dairy powders observed by X-ray photoelectron spectroscopy and effects on their rehydration properties. Colloids and Surfaces B: Biointerfaces 49, 7178. Gaiani, C., Scher, J., Ehrhardt, J.J., Linder, M., Schuck, P., Desobry, S., Banon, S., 2007. Relationships between dairy powder surface composition and wetting properties during storage: importance of residual lipids. Journal of Agricultural and Food Chemistry 55, 65616567. Gerin, P.A., Dengis, P.B., Rouxhet, P.G., 1995. Performance of XPS analysis of model biochemical compounds. Journal of Chemical Physics 92, 10431065. Hoseney, R.C., 1994. Dry milling of cereals. In: Principles of Cereal Science and Technology. American Association of Cereal Chemists, St. Paul, MN, USA, pp. 125145. Johnson, P.N.T., Brennan, J.G., 2000. Kinetics of moisture absorption by plantain our. Journal of Food Engineering 45, 3336. Ketal, L.V., Graham, B., David, C., 2004. The use of dynamic vapour sorption and near infra-red spectroscopy (DVS-NIR) to study the crystal transitions of theophylline and the report of a new solid-state transition. European Journal of Pharmaceutical Sciences 22, 97105. Kim, E.H.J., Chen, X.D., Pearce, D., 2002. Surface characterization of four industrial spray dried dairy powders in relation to chemical composition,

Identication of novel haze-active beer proteins by proteome analysis

Takashi Iimure a, *, Nami Nankaku b, Megumi Watanabe-Sugimoto c, Naohiko Hirota a, Zhou Tiansu a, Makoto Kihara a, Katsuhiro Hayashi a, Kazutoshi Ito d, Kazuhiro Sato b
a
Bioresources Research and Development Department, Sapporo Breweries Ltd., 37-1, Nittakizaki, Ota, Gunma 370-0393, Japan Barley Germplasm Center, Research Institute for Bioresources, Okayama University, 2-20-1, Chuo, Kurashiki, Okayama 710-0046, Japan c Graduate school of Natural Science and Technology, Okayama University, 1-1-1, Tsushimanaka, Okayama 700-8530, Japan d Frontier Laboratories of Value Creation, Sapporo Breweries Ltd., 10 Okatohme, Yaizu, Shizuoka 425-0013, Japan
b
Article history: Received 30 June 2008 Received in revised form 31 July 2008 Accepted 15 August 2008 Keywords: Beer colloidal haze Barley Proteome analysis
a b s t r a c t
Colloidal haze reduces beer quality considerably. Four haze samples were analyzed by two-dimensional gel electrophoresis (2DE) in order to identify haze-active proteins. Several protein spots were observed in all of the four haze samples. Using mass spectrometry analysis followed by a database search identied these spots as barley dimeric alpha-amylase inhibitor (BDAI-1), CMb component of tetrameric alphaamylase inhibitor (CMb) and trypsin inhibitor CMe precursor (CMe). These proteins were considered to be haze-active. Since haze-active proteins are adsorbed by silica gel in the beer ltration process, we eluted proteins adsorbed onto silica gel (PAS) and identied their species. These major PAS were identied as protein Z4, protein Z7 and trypsin/amylase inhibitor pUP13 (TAI), rather than BDAI-1, CMb and CMe. Furthermore, we analyzed proline compositions in the beer proteins, PAS and the haze proteins. Consequently, we found that the proline compositions of PAS were higher (ca. 20 mol%) than those in the beer proteins (ca. 10 mol%), although those of the haze-active proteins such as BDAI-1, CMb and CMe were 6.68.7 mol%. Our results suggest that BDAI-1, CMb and CMe are not predominant haze-active proteins, but growth factors of beer colloidal haze. 2008 Elsevier Ltd. All rights reserved.
1. Introduction In clear beers, colloidal storage haze is one of the principal indicators of beer quality. Consumers rely greatly on visual impressions and will generally judge a beer as stale or not t to drink if it displays haze when it is supposed to be clear. The cause of storage haze has been identied as due to interactions between haze-active proteins and certain polyphenols (Asano et al., 1982; Siebert, 1999; Siebert et al., 1996). The addition of silica gel during the beer ltration process is effective in removing haze-active proteins (Leiper et al., 2003; Siebert and Lynn, 1997). Haze-active polyphenols can be removed by the addition of polyvinylpolypyrolidone (PVPP) in beer ltration (McMurrough et al.,
Abbreviations: BDAI-1, barley dimeric alpha-amylase inhibitor; CMb, CMb component of tetrameric alpha-amylase inhibitor; CMe, trypsin inhibitor CMe precursor; 2DE, two-dimensional gel electrophoresis; LC-MS/MS, liquid chromatography mass spectrometry/mass spectrometry; LTP, lipid transfer protein; MALDI TOF-MS, matrix-assisted laser desorption/ionization time-of-ight mass spectrometry; NCBI-nr, the non-redundant amino acid database of the National Center for Biotechnology Information; PAS, proteins adsorbed onto the silica gel; TAI, trypsin/amylase inhibitor pUP13; Trx-2p, thioredoxin from yeast. * Corresponding author. Tel.: 81 276 56 1454; fax: 81 276 56 1605. E-mail address: takashi.iimure@sapporobeer.co.jp (T. Iimure). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.08.004
1995; Mikyska et al., 2002). Using barley cultivars that lack proanthocyanidins is also effective to produce beer without hazeactive polyphenols (Fukuda et al., 1999; von Wettstein et al., 1980, 1985). Although research has been conducted on haze-active proteins and polyphenols (Apperson et al., 2002; Evans et al., 2003; Leiper et al., 2003; Robinson et al., 2004), the factors controlling colloidal haze formation are still not clear. A series of beer proteins such as hordeins (Asano et al., 1982), lipid transfer protein 1 (Evans and Hejgaard, 1999; van Nierop et al., 2004) and protein Z (Evans and Hejgaard, 1999) are derived from barley malt. Hordeins are the predominant barley storage proteins and contain high levels of proline and glutamine. They are known to be involved in haze formation (Asano et al., 1982). Silica gel adsorbs proline rich proteins such as hordeins (Leiper et al., 2003). Robinson et al. (2007a,b) suggested that the barley trypsin inhibitor CMe precursor (CMe) was a haze-active protein based on the analysis of eluate proteins from silica gel. They distinguished the cultivars by the presence of CMe bands on immuno blot analysis with antibodies raised against silica eluate proteins, and revealed that beer brewed from cultivars without the CMe band formed less haze than the one with the CMe band (Evans et al., 2003; Robinson et al., 2004). Although several non-hordein proteins are regarded as haze-active, functions of these proteins in colloidal haze formation have not been well characterized. To reveal the mechanism of haze
142
T. Iimure et al. / Journal of Cereal Science 49 (2009) 141147
formation, beer proteins should be more comprehensively investigated. Proteome analysis, such as two-dimensional gel electrophoresis (2DE), followed by protein mass spectrometry, is a powerful tool to identify potential proteins of interest. Using this technique, barley grain, barley malt (Bak-Jensen et al., 2004; stergaard et al., 2002) and beer (Perrocheau et al., 2005) have been analyzed. We also analyzed the beer proteins using 2DE and identied a possible foam-promoting protein (Okada et al., 2008). In this study, we prepared four haze-positive beer samples, analyzed proteins in haze by 2DE and identied haze-active proteins. We assumed that the protein spots observed on all four haze samples were from haze-active proteins, and that any proteins not appearing on the gel were haze-inactive proteins. In addition, in order to investigate the composition of proteins adsorbed onto the silica gel (PAS), we analyzed the PAS by 2DE and compared the proline compositions of the beer proteins, PAS and the haze proteins. 2. Experimental 2.1. Beer and haze samples To prepare haze-positive beer samples, commercial beers were stored for over one year at 4 C. These haze-positive beer samples were used for the analyses. Beer brewed from the malt of North American cultivar A, as described in our previous paper (Okada et al., 2008), was used to prepare a 25% salt-precipitated protein fraction. Beer samples ltered with and without silica gel (SiO2 beer and SiO2 beer) were brewed using only malt and hop, according to the standard method of Sapporo Breweries Ltd., as described by Okada et al. (2008). To prepare SiO2 beer, the beer sample was ltered with 200 ppm of silica gel. 2.2. Beer quality analysis Beer characteristics were analyzed according to the standard methods of the European Brewery Convention (EBC Analytica, 1987). Colloidal stability was scored by a forcing test (FT-3). The bottled beer samples were stored at 60 C for 3 days, and then 0 C for 1 day. Subsequently, the beer clarity, FT-3, was assessed in terms of haze formation. 2.3. Protein sample preparation To separate the haze and the beer fractions, the haze-positive beer samples were centrifuged at 2000 g, for 30 min at 4 C. The precipitate was washed twice in 5% ethanol solution, before being dissolved in 8 M urea (Wako, Japan) 2% 3-[(3-cholamidopropyl) dimethylammonio] propansulfonic acid (CHAPS) (Dojindo Laboratories, Japan) solution containing 0.28% dithiothreitol (Wako). Three milliliter of this solution and supernatant were applied to a PD-10 column (GE Healthcare Biosciences, Japan), and the desalted proteins were eluted by 4 ml of distilled water. The protein concentration was determined by the Bradfords (1976) method using bovine serum albumin as a standard. These solutions were lyophilized and the lyophilized protein samples were subsequently used in 2DE and amino acid composition analyses. The 25% saltprecipitated protein fraction of the beer brewed from the malt of cultivar A was prepared according to Okada et al. (2008). The proteins adsorbed onto the silica gel (PAS) were prepared as described below by modication of the method described by Evans et al. (2003). Silica gel was added to the degassed beer sample upto 200 ppm, whereupon the solution was stirred for 1 h at room temperature and then centrifuged at 2000 g for 30 min at 4 C. 2% (v/v) of ammonium solution was added to the precipitate.
Subsequently, to separate the silica gel and proteins, the solution was stirred for 1 h at room temperature and centrifuged at 2000 g for 30 min at 4 C. By adding HCl, the pH of the solution was adjusted to 8.0. Subsequently, the solution was desalted using a PD-10 column (GE Healthcare Biosciences) and then lyophilized. The lyophilized protein was dened as PAS. 2.4. Two-dimensional gel electrophoresis (2DE) The 2DE of both the rst and second dimensions using the Multiphor II system (GE Healthcare Biosciences), and silver staining were carried out according to Okada et al. (2008). 2.5. Mass spectrometry analysis and protein identication Selected protein spots were excised from the 2DE gel and digested with trypsin as described in a previous paper (Okada et al., 2008). Eluted proteins were applied to matrix-assisted laser desorption/ionization time-of-ight mass spectrometry (MALDI TOF-MS) as described in Okada et al. (2008). Proteins were identied by peptide mass nger printing on the non-redundant amino acid database of the National Center for Biotechnology Information (NCBI-nr) using MASCOT software (Perkins et al., 1999). When a protein was not identied, it was digested using trypsin and/or lysil endopeptidase and the sample was analyzed using liquid chromatography mass spectrometry/mass spectrometry (LC-MS/ MS). A TrisHCl buffer (pH 8.0) containing lysil endopeptidase was added to decolorized sample gel prepared according to Okada et al. (2008). Then the gel was incubated for 3 h at 35 C. Subsequently, trypsin was added to this sample and incubated for 20 h at 35 C. Each sample solution was applied to LC-MS/MS. The condition of the LC-MS/MS analysis is described below. Equipment: MAGIC 2002 (Michrom BioResources, Inc., USA), column: Magic C18 (0.1 150 mm, Michrom BioResources, Inc.), mobile phase: 2% acetonitrile 0.1% formic acid, and 90% acetonitrile 0.1% formic acid, ow rate: 250300 nL min1, mass spectrometer: Q-Tof2 (Micromass, U.K.), ionization method: Nanoow-LC ESI, ionization mode: positive mode, electric potential of capillary: 1.8 kV, collision energy: 2056 eV. To identify the protein species, NCBI-nr was searched for using the resultant values of product ion from all precursor ions using the MASCOT search engine (Perkins et al., 1999). 2.6. Analysis of the proline composition of protein fraction The proline compositions of beer proteins, haze proteins and PAS were analyzed using the Waters Pico$Tag system (Waters, Japan). A 20 mL aliquot of 6 M HCl (1% phenol) was added to the lyophilized protein sample, and then the proteins were completely hydrolyzed at 110 C for 20 h. Subsequently, the amino acid concentration was measured. Amino acid concentrations were determined as described below. The sample was ultra-ltered by Microcon Ultracel YM-10, MW 10,000 (Nihon Millipore Ltd., Japan). Once the ltrate had been diluted to a suitable concentration, uorescence derivatization was carried out. A 20 mL aliquot of sample solution, 60 mL of AccQ$Fluorborate buffer (Waters) and 20 mL of AccQ Fluor reagent (Waters) were mixed, and the mixture was subsequently incubated for 10 min at 55 C. The conditions for high performance liquid chromatography are described as follows: equipment: 2695 separation module, detector: 2475 multi l uorescent detector (Waters) with Ex. 250 nm and Em. 395 nm, column: an AccQ Tag column (Waters), mobile phase: 100 mM sodium acetate 5.6 mM triethylamine (pH 5.7), 100 mM sodium acetate 5.6 mM triethylamine (pH 6.8), acetonitrile and distilled water. The chromatography data produced was analyzed using Empower personal software (Waters).
143
3. Results 3.1. 2DE analysis of the proteins in haze After we prepared the haze-positive beer samples, the haze and beer protein fractions were separated by centrifugation. Then the protein fractions were analyzed by 2DE (Fig. 1). On the 2DE gel of the beer proteins, a large intensely staining spot (spot b0) at about pI 4.55.5, and a Mr of 3545 kDa was observed. However, the intensity of spot b0 from the haze protein was lower than that of the beer (Fig. 1A, Table 1). In the haze proteins, intense spots were observed in region I-I, b810, b13, and b14 (Fig. 1). Spots in region I-I and b8 were intense both in the beer and the haze proteins. On the other hand, spots b9, b10, b13 and b14 were only intense in the haze proteins, and spots b11 and b12 were only intense in the beer proteins (Fig. 1, Table 1). The low molecular weight region of the 2DE image of the 25% salt-precipitated protein fraction was similar to those of the haze proteins, but different from those of the beer proteins (Fig. 1, Table 1). The spots of the 25% salt-precipitated fraction were distinct in region I-I, b0 and b8b10, but faint or invisible in b5 and b11b14. Fig. 2 shows the 2DE images in region I of the four haze protein samples. The protein spots were classied into three groups i.e. (1) haze-active, because the protein spot was observed in all four haze samples, (2) moderately haze-active because the protein spot was observed in few samples as a relatively faint spot or observed in all samples but as a faint spot, and (3) haze-inactive because the protein spot was faint or invisible in all four samples. Protein spots in region I-I, b10, b13 and b14 were observed in all four samples. Conversely, protein spots of b11 and b12 were invisible, and b0 (data not shown) was faint in all four samples. In addition, spot b5 was only observed in sample i faintly, and spots b8 and b9 were observed in 2 and 3 samples respectively (Table 1, Fig. 2). Overall, the protein spots are: (1) haze-active in
region I-I, b10, b13 and b14, (2) moderately haze-active in spots b0, b8 and b9, and (3) haze-inactive in spots b5, b11 and b12, respectively (see also Table 1). 3.2. Mass spectrometry analysis and protein identication To identify protein species, we analyzed the protein spots in region I-I, b0, b5, and b8b14 by MALDI TOF-MS or LC-MS/MS followed by a database search. Table 2 shows the identied protein spots by the analyses. Considering the classication of regions by haze-activity, barley dimeric alpha-amylase inhibitor (BDAI-1), CMb component of tetrameric alpha-amylase inhibitor (CMb), and trypsin inhibitor CMe precursor (CMe) were found in haze-active regions. Protein Z-type serpin (protein Z4), serpin (protein Z7), and chain A, non-specic lipid transfer protein 1 (LTP) were found in moderately haze-active regions. Thioredoxin from yeast (Trx-2p) and trypsin/amylase inhibitor pUP13 (TAI) were found in hazeinactive regions. 3.3. 2DE analysis of the proteins adsorbed onto the silica gel (PAS) Silica gel is known to adsorb haze-active proteins (Leiper et al., 2003; Siebert and Lynn, 1997) therefore, an addition of silica gel at beer ltration is effective to prevent beer colloidal haze. We prepared two beer samples with and without silica gel in ltration, described as SiO2 beer and SiO2 beer, respectively. The quality proles of these beer samples were similar except for FT-3 i.e. the beer clarity after the forcing test (Table 3). The SiO2 beer showed resistance to beer colloidal haze. On the other hand, the SiO2 beer was shown to be prone to haze formation in response to force testing and contained haze-active protein, which was partly adsorbed onto the silica gel. To prepare the proteins adsorbed onto the silica gel (PAS), we added silica gel to the SiO2 beer, and then eluted proteins
Mr (kDa) 97.4 66.2 45.0 31.0 21.5 14.4 97.4 66.2 45.0 31.0 21.5 14.4
pI3
pI 10 Beer (supernatant) b0
B
Mw (kDa) 14.4 b11 b12 b8 Region I-I b5 b2
Region I
Haze (precipitate) b9 14.4 b10 b8 b13 b14
97.4 66.2 45.0
25% salt precipitate
14.4 31.0 21.5 14.4
Fig. 1. The two-dimensional gel electrophoresis (2DE) images of the beer proteins of sample i, the haze proteins of sample i and the 25% salt-precipitated proteins of the beer brewed from the malt of cultivar A. A: The whole images of 2DE, pI 310, B: the enlarged images of region I in A. Arrows indicate the spot numbers.
144
Table 1 Summary of the spot intensities of region I-I, b0, b5 and b8b14 in the beers, the hazes, the 25% salt-precipitated proteins and the proteins adsorbed onto the silica gel (PAS). The beer proteins The haze proteins Sample i b0 Region I-I b5 b8 b9 b10 b11 b12 b13 b14
a b c
Sample ii
Sample iii
Sample iv
The 25% salt-precipitated proteins

B B
Haze-activity
PASa
B B
6
B
6
B
6
B
6
B
6c
B
6
B
6
B B B
6
B

B B
6 6

B B B
6 6
B
6

B B
6 6 6
B B

B B
6 6
6
B

B B
PAS, proteins adsorbed onto the silica gel. B indicates that the spot was distinct, 6 indicates that the spot was faint, and indicates that the spot was invisible. B indicates that the protein was haze-active because the protein spot was observed in all four haze samples, 6 indicates that the protein was moderately haze-active because the protein spot was observed in few samples as relatively faint spot or observed in all samples but as faint spot, and indicates that the protein was haze-inactive because the protein spot was faint or invisible in all four samples.
from the silica gel. Subsequently, we analyzed the proteins of two beer samples and PAS. The 2DE images of the two beer proteins were quite similar (Fig. 3). The 2DE images in PAS and the haze proteins were different both at the intensity of the spot b0 (protein Z) (Figs.1A and 4A) and the region I in the low molecular region (Figs. 2 and 4). The intensity of spot b0 in PAS was higher than those in the haze
sample (Figs. 1A and 4A). In the PAS at region I (Fig. 4B), intensities of spots b8 (CMb), b9 (LTP), b10 (CMb), b13 (CMe) and b14 (CMe) were faint or invisible, those in region I-I (BDAI-1) were low except for the spot b2, and those in b11 and b12 (TAI) were distinct, respectively. These results suggest that the major PAS are protein Z4, protein Z7, and TAI, rather than BDAI-1, CMb and CMe.
Fig. 2. The two-dimensional gel electrophoresis (2DE) images of the low molecular region, region I (Fig. 1A) of the haze samples i, ii, iii and iv. Arrows indicate the spot numbers.
T. Iimure et al. / Journal of Cereal Science 49 (2009) 141147 Table 2 Summary of protein identication by mass spectrometry analysis followed by a database search. Protein spot number b0 Region I-Ia Protein name protein Z-type serpin serpin barley dimeric alpha-amylase inhibitor; BDAI-1 thioredoxin; Trx-2p CMb component of tetrameric alpha-amylase inhibitor chain A, non-specic lipid transfer protein 1 CMb component of tetrameric alpha-amylase inhibitor trypsin/amylase inhibitor pUP13 trypsin/amylase inhibitor pUP13 trypsin inhibitor CMe precursor trypsin inhibitor CMe precursor Accession number CAA66232 CAA64599 CAA08836 Organism barley barley barley
145
hand, the proline compositions of the two haze proteins (samples i and iii) differed considerably (Table 4). 4. Discussion To identify haze-active proteins, we prepared four haze-positive beer samples and analyzed the haze proteins using 2DE and mass spectrometry followed by a database search. Consequently, we assume that BDAI-1, CMb and CMe are haze-active, protein Z4, protein Z7 and LTP are moderately haze-active, and Trx-2p and TAI are haze-inactive proteins, respectively. Among the haze-active proteins the spots b10 (CMb), b13 (CMe) and b14 (CMe) were not observed in the beer protein sample i (Fig. 1) but spots in region I-I (BDAI-1) were intense (Fig. 1B). It is suggested, therefore, that CMb and CMe are concentrated in the haze fraction. In salting-out of the beer proteins, proteins did not precipitate at up to 20% saturation, but precipitated at 25% saturation of ammonium sulfate (data not shown). Therefore, the 25% salt-precipitated protein fraction may contain relatively hydrophobic proteins. The intense protein spots in region I-I (BDAI-1), b8 and b10 (CMb) were observed in a 25% salt-precipitated fraction (Fig. 1B), suggesting that BDAI-1 and CMb are potentially haze-active due to their high hydrophobicity. CMe is a possible haze-active protein although the mechanism is unknown at present. The spot intensities of regions I-I (BDAI-1), b10 (CMb), b13 and b14 (CMe) were different among the four haze samples (Fig. 2). In haze sample ii in Fig. 2 the protein spots b10 (CMb), b13 and b14 (CMe) were less prominent despite the high intensity of the protein spots in region I-I (BDAI-1). This could have been due to the difference in the level of haze formability and suggested that BDAI1 was more haze-active than both CMb and CMe. From the analyses of our four haze-positive beer samples, we consider that BDAI-1, CMb and CMe contribute to haze formation. The association of CMe with haze formation has already been suggested by Evans et al. (2003) and Robinson et al. (2004). They indicated that beer brewed from the malt of several barley varieties without the CMe band in western blot analysis using the antibody of silica eluate proteins showed improved haze stability compared to those with the CMe band. In order to further conrm the function of other haze-active proteins, i.e. BDAI-1 and CMb, we need to perform further experiments; for example force tests where these proteins are added to beer, and/or the development of novel cultivars without BDAI-1 or CMb which are then subject to the same trials. It has been shown that the addition of silica gel during beer ltration contributes to haze reduction due to adsorption of hazeactive proteins onto silica gel (Leiper et al., 2003; Siebert and Lynn, 1997). We prepared two beer samples that were ltered with and without silica gel, i.e. SiO2 beer and SiO2 beer, respectively. The
b5 b8
NP_011725 CAA49556
yeast barley
b9 b10
1MID-A CAA49556
barley barley
b11 b12 b13 b14

a
1208404A 1208404A P01086 P01086
barley barley barley barley
All spots in region I-I.
Table 3 The characteristics of the beer samples ltrated with silica gel (SiO2 beer) and ltrated without silica gel (SiO2 beer). SiO2 beer original gravity (%) nal extract (%) apparent extract (%) alcohol (vol.%) pH color (EBC) bitter unit (mg/L) FT-3a
a
SiO2 beer 11.49 3.88 2.07 4.99 4.50 7.6 28.1 1.68
11.67 3.95 2.11 5.07 4.49 7.9 28.2 >10.0
FT-3 means the beer clarity after the forcing test (see Section 2).
3.4. Proline compositions of the beer, the haze and the PAS It is well known that hordeins, one of the haze-active proteins, are proline rich (Asano et al., 1982). We examined the proline compositions of the beer proteins, the haze proteins and PAS (Table 4). The proline compositions in the beer proteins were ca. 10 mol%, while those in PAS were ca. 20 mol%. These results suggest that proline rich proteins are adsorbed onto the silica gel. On the other
Mr (kDa) 97.4 66.2 45.0 31.0 21.5 14.4
pI 3
pI 10 SiO2+ beer
pI 3
pI 10 SiO2-beer
Fig. 3. The two-dimensional gel electrophoresis images of the beer ltered with silica gel (SiO2 beer) and of that ltered without silica gel (SiO2 beer).
146
Mr (kDa) 97.4
pI 3
pI 10
PASa
66.2 45.0 31.0
21.5 14.4
Region I
B
b11 Region I-I
b9 b12 b10 b8 b13
PAS
b14
b2
Fig. 4. The two-dimensional gel electrophoresis (2DE) images of the proteins adsorbed onto the silica gel (PAS). A: The whole image of the 2DE, pI 310, B: the enlarged image of region I. Arrows indicate the spot numbers.
It is conventionally accepted that haze proteins have elevated levels of proline (Siebert and Lynn, 1997). The proline composition of PAS was ca. 20 mol% (Table 4), while those of BDAI-1, CMb and CMe were as low as 6.6, 8.7 and 6.8 mol%, respectively. On the other hand, the proline composition of the haze proteins varied considerably between samples i (5.6 mol%) and iii (27.9 mol%), respectively (Table 4). If the beer colloidal haze consists of only proline rich proteins, such as hordein, the proline composition in all haze samples would be expected to be high (ca. 20 mol%). In addition, if the beer colloidal haze consists of only relatively proline poor proteins, such as BDAI-1, CMb and CMe, the proline compositions in all haze samples would be low (69 mol%). We estimate, therefore, that the proline composition of colloidal haze proteins might be determined by the degree of interaction between proline poor proteins, such as BDAI-1, CMb and CMe, and a core colloidal haze consisting of proline rich proteins such as hordeins. It is suggested that BDAI-1, CMb and CMe are not predominant haze-active proteins, but in fact initiation or growth factors in the formation of colloidal haze. Proline rich protein such as hordein was not detected in the 2DE analysis of haze proteins, because hordein may not have been detectable by the 2DE analysis applied in this study due to being poorly silver stained. Therefore, our results do not exclude the contribution of hordein-derived polypeptides to haze formation. To further improve the beer colloidal haze, we need to develop novel silica gel having adsorbing afnity to BDAI-1, CMb and CMe or develop cultivars with low haze-active proteins. Okada et al. (2008) suggested that BDAI-1 is a possible foam-positive protein. Therefore the amount of BDAI-1 may well need to be optimized to produce beer with high qualities such as foam retention and haze stability. Acknowledgements
haze stability of SiO2 beer was improved (Table 3). We then analyzed PAS using 2DE and identied protein species using mass spectrometry followed by a database search. As a result, the major proteins adsorbed onto the silica gel were identied as protein Z4, protein Z7 and TAI, rather than BDAI-1, CMb and CMe. Therefore, the majority of BDAI-1, CMb and CMe were transferred to beer after ltration. If BDAI-1, CMb and CMe were the predominant hazeactive proteins, the level of haze stability in SiO2 beer might be comparable with that of SiO2 beer. However, the haze stability in SiO2 beer was improved compared to SiO2 beer (Table 3). This evidence demonstrates that current silica treatment practices are only partially effective in improving haze stability because the majority of the BDAI-1, CMb and CMe remain in the treated beer. It follows that if these proteins could be removed, either by selection of novel barley varieties that do not contain these proteins, or improved haze stability treatments, the colloidal stability of beer could be further improved.
We are grateful to T. Yazawa, K. Ito and H. Kato of the Bioresources Research and Development Department, Sapporo Breweries Ltd. for their technical assistance. This study was supported by the Program for Promotion of Basic Research Activities for Innovative Biosciences, Japan (PROBRAIN). References
Apperson, K., Leiper, K.A., McKeown, I.P., Birch, D.J.S., 2002. Beer uorescence and the isolation, characterization and silica adsorption of haze-active beer proteins. Journal of the Institute of Brewing 108, 193199. Asano, K., Shinagawa, K., Hashimoto, N., 1982. Characterization of haze-forming proteins of beer and their roles in chill haze formation. Journal of the American Society of Brewing Chemists 40, 147154. Bak-Jensen, K.S., Laugesen, S., Roepstorff, P., Svensson, B., 2004. Two-dimensional gel electrophoresis pattern (pH 611) and identication of water-soluble barley seed and malt proteins by mass spectrometry. Proteomics 4, 728742. Bradford, M.M., 1976. A rapid and sensitive method for the quantization of microgram quantities of protein utilizing the principle of proteindye binding. Analytical Biochemistry 72, 248254. European Brewery Convention, Analytica, 1987. European Brewery Convention, fourth ed. Brauerei und Getraenke Rundschau, Zurich. Evans, D.E., Hejgaard, J., 1999. The impact of malt derived proteins on beer foam quality part I: the effect of germination and kilning on the level of proteins Z4, protein Z7 and LTP1. Journal of the Institute of Brewing 105, 159169. Evans, D.E., Robinson, L.H., Sheehan, M.C., Tolhurst, R.L., Hill, A., Skerritt, J.S., Barr, A.R., 2003. Application of immunological methods to differentiate between foam-positive and haze-active proteins originating from malt. Journal of the American Society of Brewing Chemists 61, 5562. Fukuda, K., Saito, W., Arai, S., Aida, Y., 1999. Production of a novel proanthocyanidin-free barley line with high quality. Journal of the Institute of Brewing 105, 179183. Leiper, K.A., Stewart, G.G., McKeown, I.P., 2003. Beer polypeptides and silica gel part I. Polypeptides involved in haze formation. Journal of the Institute of Brewing 109, 5772. McMurrough, I., Madigan, D., Smyth, M.R., 1995. Adsorption by polyvinylpolypyrrolidone of catechins and proanthocyanidins from barley. Journal of Agricultural and Food Chemistry 43, 26872691.
Table 4 The proline compositions of the beer proteins, the proteins adsorbed onto the silica gel (PAS) and the haze proteins. Sample name Beer PASa Haze
a
Proline (mol%) 11.4 10.7 16.2 23.4 5.6 27.9
SiO2 beer SiO2 beer SiO2 beer SiO2 beer sample i sample iii
PAS, proteins adsorbed onto the silica gel.
T. Iimure et al. / Journal of Cereal Science 49 (2009) 141147 Mikyska, A., Hrabak, M., Haskova, D., Srogl, J., 2002. The role of malt and hop polyphenols in beer quality, avour and haze stability. Journal of the Institute of Brewing 108, 7885. Okada, Y., Iimure, T., Takoi, K., Kaneko, T., Kihara, M., Hayashi, K., Ito, K., Sato, K., Takeda, K., 2008. The inuence of barley malt protein modication on beer foam stability and their relationship to the barley dimeric a-amylase inhibitor-I (BDAI-I) as a possible foam-promoting protein. Journal of Agricultural and Food Chemistry 56, 14581464. stergaard, O., Melchior, S., Roepstorff, P., Svensson, B., 2002. Initial proteome analysis of mature barley seeds and malt. Proteomics 2, 733739. Perkins, D.N., Pappin, D.J., Creasy, D.M., Cottrell, J.S., 1999. Probability-based protein identication by searching sequence databases using mass spectrometry data. Electrophoresis 20, 35513567. Perrocheau, L., Rogniaux, H., Boivin, P., Marion, D., 2005. Probing heat-stable watersoluble proteins from barley to malt and beer. Proteomics 5, 28492858. Robinson, L.H., Evans, D.E., Kaukovirta-Norja, A., Vilpola, A., Aldred, P., Home, S., 2004. The interaction between malt protein quality and brewing conditions and their impact on beer colloidal stability. Technical Quarterly Master Brewers Association of the Americas 41, 353362. Robinson, L.H., Healy, P., Stewart, D.C., Eglinton, J.K., Ford, C.M., Evans, D.E., 2007a. The identication of a barley haze active protein that inuences beer haze stability: the genetic basis of a barley malt haze active protein. Journal of Cereal Science 45, 335342.
147
Robinson, L.H., Juttner, J., Milligan, A., Lahnstein, J., Eglinton, J.K., Evans, D.E., 2007b. The identication of a barley haze active protein that inuences beer haze stability: cloning and characterization of the barley SE protein as a barley trypsin inhibitor of the chloroform/methanol type. Journal of Cereal Science 45, 343352. Siebert, K.J., Carrasco, A., Lynn, P.Y., 1996. Formation of proteinpolyphenol haze in beverages. Journal of Agricultural and Food Chemistry 44, 19972005. Siebert, K.J., Lynn, P.Y., 1997. Mechanisms of beer colloidal stabilization. Journal of the American Society of Brewing Chemists 55, 7378. Siebert, K.J., 1999. Effects of proteinpolyphenol interactions on beverage haze, stabilization, and analysis. Journal of Agricultural and Food Chemistry 47, 353362. van Nierop, S.N.E., Evans, D.E., Axcell, B.C., Cantrell, I.C., Rautenbach, M., 2004. Impact of different wort boiling temperatures on the beer foam stabilizing properties of lipid transfer protein 1. Journal of Agricultural and Food Chemistry 52, 31203129. von Wettstein, D., Jende-Strid, B., Ahrenst-Larsen, B., Erdal, K., 1980. Proanthocyanidin-free barley prevents the formation of beer haze. Technical Quarterly Master Brewers Association of the Americas 17, 1623. von Wettstein, D., Nilan, R.A., Ahrenst-Larsen, B., Erdal, K., Ingversen, J., JendeStrid, B., Kristiansen, K.N., Larsen, J., Outtrup, H., Ullrich, S.E., 1985. Proanthocyanidin-free barley for brewing: progress in breeding for high yield and research tool in polyphenol chemistry. Technical Quarterly Master Brewers Association of the Americas 22, 4152.

Mixing properties and dough functionality of transgenic lines of a commercial wheat cultivar expressing the 1Ax1, 1Dx5 and 1Dy10 HMW glutenin subunit genes
Elena Leon a, Santiago Marn a, Mara J. Gimenez a, Fernando Piston a, Marta Rodrguez-Quijano b, Peter R. Shewry c, Francisco Barro a, *
a b c
Instituto de Agricultura Sostenible, CSIC, Alameda del Obispo s/n, 14080 Cordoba, Spain Universidad Politecnica de Madrid, 28040 Madrid, Spain Rothamsted Research, Harpenden, Hertfordshire AL5 2JQ, UK
Article history: Received 18 April 2008 Received in revised form 13 August 2008 Accepted 18 August 2008 Keywords: 1Ax1 1Dx5 1Dy10 Anza GM wheat
a b s t r a c t
In this work we report the effects of the HMW-GS 1Ax1, 1Dx5 and 1Dy10 on the breadmaking quality of the bread wheat cultivar Anza that contains the HMW-GS pairs 1Dx2 1Dy12 and 1Bx7* 1By8, and is null for the Glu-A1 locus. This allows the characterization of individual subunits 1Dx5 and 1Dy10 in the absence of subunit 1Dx5, and the interactions between these subunits and subunits 1Dx2 and 1Dy12 to be determined. Three transgenic lines termed T580, T581 and T590, containing, respectively, the HMWGS 1Ax1, 1Dx5 and 1Dy10 were characterized over 3 years using a range of widely-used grain and dough testing methods. The transgenic subunits 1Ax1, 1Dx5 and 1Dy10 accounted for 25.2%, 20.3% and 17.9%, respectively, of the total HMW-GS in the three transgenic lines. Although lines T581 and T590 expressed similar levels of subunits 1Dx5 and 1Dy10 they had different effects on other aspects of protein composition, including changes in the ratios of glutenin/gliadin, of HMW/LMW-GS, the 1Dx2/1Dy12, the x-type/y-type HMW-GS and the proportions of high molecular mass glutenin polymers. In contrast, lines transformed to express subunits 1Ax1 and 1Dx5 showed similar changes in protein composition, with higher protein contents and decreased ratios of glutenin/gliadin and 1Dx2/1Dy12. In addition, both transgenic lines showed similar increases in the ratio of x-type/y-type subunits compared to the control line. The transgenic lines were analysed using Farinograph, Mixograph and Alveograph. This conrmed that the expression of all three subunits resulted in increased dough strength (and hence breadmaking quality) of the cultivar Anza. A benecial effect of subunit 1Dx5 has not been reported previously, transgenic wheat lines expressing this subunit giving overstrong dough unsuitable for breadmaking. However, the expression of subunit 1Dy10 had a greater effect on breadmaking quality than subunits 1Ax1 and 1Dx5. The Farinograph parameters such as dough stability and peak time were increased by 9.2-fold and 2.4-fold, respectively, in line T590 (expressing 1Dy10) with respect to the control line. Similarly, the Mixograph mixing time was increased by four-fold and the resistance breakdown decreased by two-fold in line T590 compared with the control line. The Alveograph W value was also increased by 2.7-fold in line T590 compared to the control line. These transgenic lines are of value for studying the contribution of specic HMW-GS to wheat our functional properties. 2008 Elsevier Ltd. All rights reserved.
1. Introduction The viscoelastic properties of wheat our dough allow wheat to be used for making bread and many other food products such as cake, biscuits, pasta and noodles. The unique properties of wheat result from the unusual biomechanical properties of the gluten proteins, which form a network conferring elasticity and
* Corresponding author. Tel.: 34 957 499240; fax: 34 957 499252. E-mail address: ge1balof@uco.es (F. Barro). 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.08.002
extensibility to the dough. One group of gluten proteins, the high molecular weight glutenin subunits (HMW-GS), play an important role in determining the functional properties of wheat dough. The HMW-GS of wheat have been studied in detail, demonstrating that allelic differences in their composition result in effects on glutenin polymers and hence breadmaking quality (Payne, 1987; Shewry et al., 2003). Bread wheat contains six HMW-GS genes, with tightly linked pairs of genes encoding x- and y- type subunits being present at each of the Glu-A1, Glu-B1, and Glu-D1 loci on the long arms of chromosomes 1A, 1B, and 1D, respectively. Wheat cultivars
E. Leon et al. / Journal of Cereal Science 49 (2009) 148156
149
containing HMW-GS 1Dx5 and 1Dy10 have stronger doughs than those containing the allelic subunits 1Dx2 and 1Dy12 (Shewry et al., 2003). However, because the x- and y-type genes are so tightly linked in the D genome, it has not been possible to separate them by recombination and hence to determine their individual contributions to the breadmaking quality of wheat. Genetic engineering provides an opportunity to develop cultivars with new HMW-GS combinations in order to explore the molecular and biochemical bases for the breadmaking quality of wheat and to generate lines with improved or novel functional properties. This approach has also been used to investigate the effects of HMW-GS genes when expressed in other cereals. Thus, genes encoding HMW-GS have been expressed in wheat (Altpeter et al., 1996; Barro et al., 1997; Blechl and Anderson, 1996), tritordeum (Rooke et al., 1999a), maize (Sangtong et al., 2002), and rye (Altpeter et al., 2004). Expression of additional HMW-GS genes in wheat has resulted in increased amounts of HMW-GS protein and changes in the grain functional properties, the effects varying depending on the specic genes and genetic backgrounds. Thus, expression of subunit 1Ax1 in transgenic wheat resulted in lines with improved rheological properties while lines expressing subunit 1Dx5 resulted in unsuitable properties for breadmaking (Barro et al., 2003; Darlington et al., 2003). In particular, high levels of expression of subunit 1Dx5 resulted in greatly increased dough strength, being too strong for conventional breadmaking (Alvarez et al., 2001; Darlington et al., 2003; Rooke et al., 1999b). The differences between the results obtained with subunits 1Ax1 and 1Dx5 may be due to the fact that subunit 1Dx5 has an additional cysteine residue which promotes a more highly cross-linked glutenin network (Darlington et al., 2003; Popineau et al., 2001). Also, because subunit 1Dx5 only occurs with 1Dy10 it has been suggested that the effects might be mitigated by increasing subunit 1Dy10 at the same time (Barro et al., 2003; Darlington et al., 2003). More recently, the effects of expression of transgenic subunits 1Dx5 and/or 1Dy10 on our properties in a wheat line which already expresses the endogenous forms of these subunits have been reported (Blechl et al., 2007). The results conrmed the effects of the increased levels of subunit 1Dx5 on dough strength but showed that increases in subunit 1Dy10 alone also resulted in curves typical of strong doughs. Furthermore, the expression of subunits 1Dx5 and 1Dy10 in the same lines failed to alleviate the overstrong characteristics associated with subunit 1Dx5 (Blechl et al., 2007). However, the effects of subunits 1Dx5 and 1Dy10 have not been tested in wheat backgrounds that do not contain these subunits. We report here the transformation of wheat with subunits 1Ax1, 1Dx5 and 1Dy10 using a genetic background which does not express endogenous forms of subunits 1Dx5 and 1Dy10 but the allelic subunits 1Dx2 and 1Dy12. This allows the effects of individual subunits 1Dx5 and 1Dy10, and the interactions between these and subunits 1Dx2 and 1Dy12 to be determined.
2.2. Plant transformation and in vitro culture Immature scutella (0.51.5 mm in length) of wheat were used as target tissue for transformation by particle bombardment. Caryopses were harvested 16 days after anthesis and explants isolated as described by Barcelo and Lazzeri (1995). For each bombardment, 30 scutella were excised and placed in the centre of a 9 cm Petri dish containing MP4 medium which consisted of MS (Murashige and Skoog, 1962) induction medium supplemented with 30 g l1 sucrose and 4 mg l1 picloram. Explants were cultured in the dark at 25 C for 5 days prior to bombardment. Explants were subjected to a 4 h osmotic treatment, both before and after bombardment, in the same induction medium described above, but supplemented with 0.4 M mannitol. For bombardment, plasmids were precipitated onto 0.6 mm gold particles at 0.5 pmol mg1 gold for pAHC25 and 0.75 pmol mg1 gold for the plasmids containing the HMW glutenin subunit genes. For each shot, 58 mg of coated particles were delivered at a pressure of 1100 psi using a PDS 1000/He particle gun (BioRad). Following bombardment, plates containing explants were cultured in MP4 medium (see above) in the dark at 25 C for 3 weeks and then transferred to a RZ regeneration medium (Barro et al., 1998) supplemented with 2 mg l1 of PPT and cultured for a further 3 weeks at 25 C with a photoperiod of 12 h. Surviving explants were transferred to R regeneration medium (Barro et al., 1998) supplemented with 2 mg l1 PPT and cultured for a further two rounds of 3 weeks each. Putative transgenic plants were then transferred to soil and grown to maturity in the greenhouse. Homozygous progeny of plants containing HMW-GS transgenes were identied by SDSPAGE of endosperm proteins by single half-seed descent (see below). Homozygous lines were self-pollinated for three generations and assayed during 3 years, using a randomized complete block design with two replicates, as described (Barro et al., 2002). 2.3. Genomic Southern blot analysis Total genomic DNA was extracted from leaf tissue using a CTAB method (Stacey and Isaac, 1994). Genomic DNA was digested with either EcoRV (which cuts once within the p1Ax1 and pK-Dy10A constructs) or ScaI (which cuts once within the p1Dx5 construct). Digested DNA was separated by electrophoresis using 0.8% (w/v) agarose gel, blotted onto Hybond-N membrane (Amersham Biosciences) and baked for 2 h at 80 C. Hybridization and signal detection were performed according to Roche Diagnostic protocols. Briey, the membrane was prehybridized for 1 h with DIG Easy Hyb Granules (Roche Diagnostics) at 48 C. Southern lters were hybridized overnight at 48 C with PCR-generated digoxigeninlabelled probes produced using primers for the 1Ax1 (Rooke et al., 1999a), 1Dx5 (DOvidio and Anderson, 1994) and 1Dy10 genes (50 CCA CAA CAC CGA GCA CCA CAA-30 , 50 -GGG CGG CAC CAC AGT TTG CTC-30 ), with p1Ax1, p1Dx5 and pK-Dy10A as the DNA template, respectively. After hybridization the membrane was washed twice with a solution of 0.1% (w/v) SDS in 2 SSC for 15 min at 65 C. The DIG Luminescent Detection Kit (Roche Diagnostics) was used for signal detection. Finally, the membrane was exposed to AGFA medical X-ray (Agfa-Gevaert NV) lm for 16 h. 2.4. Protein and SDSPAGE analysis
2. Material and methods 2.1. Plasmids Three plasmids, p1Dx5, p1Ax1 and pK-Dy10A containing, respectively, the genomic fragment for the HMW glutenin subunits 1Dx5 (Anderson et al., 1989), 1Ax1 (Halford et al., 1992), and 1Dy10 (Anderson et al., 1989), were used in combination with plasmid pACH25 (Christensen and Quail, 1996) which contains the bar gene that confers tolerance to phosphinothricin (PPT) under the control of maize ubiquitin promoter.
The protein content of our was calculated from the Kjeldahl nitrogen content (%N 5.7) and expressed on a dry matter basis. For protein extraction and quantication of the different protein fractions, all subplots in a year from the same line were bulked. Therefore, 30 individual seeds per line (ten seeds from each year) were crushed into a ne powder and used to extract the endosperm
150
storage proteins. Gliadins were extracted in 60% (v/v) of aqueous ethanol using a rotary shaker for 40 min. Samples were centrifuged at 13,000 g for 5 min and the supernatant collected. Glutenins were extracted in 625 mM TrisHCl pH 6.8, 5% (v/v) 2-mercaptoethanol, 10% (v/v) glycerol, 0.02% (w/v) bromophenol blue and 2% (w/v) dithiothreitol in a 5:1 ratio (ml:mg) to wholemeal and separated using a Trisborate buffer system and 10% (w/v) acrylamide gels (Shewry et al., 1995). Gliadin and glutenin contents were determined in the above extracts using a Bradford assay system (Bradford, 1976). For densitometry analysis, glutenins were separated by SDSPAGE gels and analysed using a Kodak Image Station 440CF and Kodak 1D Image Analysis Software using the SDSPAGE Molecular Weight Standards from Bio-Rad as reference. Falling Number was determined according to the standard ICC method no. 107 (ICC, 1995). Sedimentation volume was determined according to Zeleny using the standard ICC method no. 116 (ICC, 1994). SE-HPLC was carried out in duplicate at the Campden and Chorleywood Food Research Association (Chipping Campden, UK) as described by Morel et al (2000) and Millar (2003). 2.5. Alveograph Approximately 0.8 kg samples of seeds per line and year were milled with a Chopin CD1 mill to obtain white our. Alveograph tests were performed using an Alveograph MA 82 (Tripette et Renaud, France) and the dough properties determined in ve replicates according to standard ICC method no. 121 (ICC, 1992a). The alveograms were evaluated to determine overpressure (tenacity) (P), the average length of the curve (extensibility) (L), the P/L ratio, the deformation work (W), the swelling index (G) and the elasticity index. 2.6. Farinograph Approximately 0.2 kg samples of seeds per line and year were milled and mixed using a 50 g mixer for standard Farinograph analysis (Brabender GmbH & Co. KG, Germany) test. The Farinograph characteristics were determined in two different replicates per year according to standard ICC method no. 115/1 (ICC, 1992b). Farinograph parameters evaluated included water absorption, arrival time, departure time, stability, peak time, mixing tolerance time and degree of softening. 2.7. Mixograph Dough mixing properties were determined with a 10 g Mixograph (National Manufacturing Co., Lincoln NE). Samples were mixed to optimum water absorption following 54-40A method (AACC, 2000). The mixing parameters determined were mixing time, peak resistance and resistance breakdown. 2.8. Statistics Data were analysed using the SPSS version 11.0 statistical software package (SPSS Inc., Chicago, IL, USA). The general analysis of variance and the least signicant difference pairwise comparisons of means were used to determine signicant differences. 3. Results Plasmids p1Ax1, p1DX5 and pKS-Dy10A, containing, respectively, the HMW-GS 1Ax1, 1Dx5 and 1Dy10, were used for transformation of the bread wheat cultivar Anza. Transformation efciency (transformants per scutellum bombarded) for this cultivar was 0.4%. This cultivar contains four endogenous HMW-GS:
1Bx7*, 1By8, 1Dx2 and 1Dy12 (Fig. 1A). Single seed descent was used to obtain homozygotes for lines containing HMW glutenin subunits transgenes by screening 30 half-seeds in each generation by SDSPAGE. Homozygous transgenic lines expressing the HMWGS 1Ax1, 1Dx5 and 1Dy10 transgenes (Fig. 1A) were termed T580, T581 and T590, respectively. These lines were multiplied by selfpollinating for three generations and assayed for 3 years as described (Barro et al., 2002). Southern blot analysis was complicated by the presence of cross-hybridizing bands in the control line, which was probably due to the sequence similarity between the endogenous and transgenic HMW-GS. Nevertheless, the banding pattern after hybridization showed that transgenic lines contained four, two and three insertion events for HMW-GS 1Ax1, 1Dx5 and 1Dy10, respectively (Fig. 1B). 3.1. Protein characterization The total protein contents of the ours ranged from 11.7% to 13.7% for the control line and line T581 transformed with HMW-GS 1Dx5, respectively (Table 1). All three transgenic lines had higher protein contents than the control line but only the differences between lines T580 and T581, containing HMW-GS 1Ax1 and 1Dx5, respectively, were statistically signicant in comparison to the control line (Table 1). As shown in Table 1, the glutenin fraction decreased from 4.21 mg/mg our in the control line to 3.25 and 3.62 in lines T580 and T581, but increased to 5.28 in line T590. In contrast, the gliadin contents increased signicantly in all three transgenic lines with respect to the control line, with no signicant differences among the three transgenic lines. These changes are also reected in the glutenin/gliadin ratio. Thus, lines T580 and T581 had lower glutenin/gliadin ratios than the control line, but line T590 had a similar ratio to that of control line (Table 1). The expression of HMW-GS in the transgenic lines resulted in changes in amounts and proportions of HMW-GS in the endosperm. Thus, the transgenic subunits 1Ax1, 1Dx5 and 1Dy10 accounted for 25.2%, 20.3% and 17.9%, respectively, of the total HMW in the three transgenic lines (Table 1). These increases in HMW subunit content were associated with decreases in LMW subunit content in all transgenic lines, but the differences were only signicant for line T581 expressing the 1Ax1 subunit (Table 1). Expression of the transgenic HMW-GS was also associated with a signicant decrease in the amounts of endogenous HMW-GS (Table 1). Furthermore, although all endogenous subunits were decreased, the relative effects varied. The expression of subunits 1Ax1 and 1Dx5 in lines T580 and T581, respectively, led to greater decreases in the contents of the endogenous subunit 1Dx2 (Table 1), with the Dx2/Dy12 ratio decreasing signicantly from 1.11 in the control line to 0.83 and to 0.88 in lines T580 and T581, respectively. Furthermore, the ratio of x-type/y-type HMW-GS also increased in lines T580 and T581, respectively. In contrast, the contents of the endogenous subunits 1Dx2 and 1Dy12 were decreased in line T590 expressing subunit 1Dy10 but to a similar extent (Table 1). Therefore, the ratio of Dx2/Dy12 in this line was comparable to that in the control line (Table 1), although the ratio of the x-type/y-type was decreased from 1.25 in the control line to 0.91 as a consequence of the addition of a new y-type subunit. There is a well-established correlation between the breadmaking performance of our and the proportion of high molecular mass glutenin polymers (Gupta and MacRitchie, 1994; Popineau et al., 1994). The proportions of monomeric, oligomeric and polymeric gluten proteins in the ours were therefore determined by SE-HPLC essentially as described by Morel et al. (2000). This method uses sonication in 1% (w/v) SDS in 0.1 M phosphate buffer, pH 6.9, to extract the total grain proteins, which are then separated into ve fractions corresponding broadly to high molecular mass glutenin
151
A
1Ax1
T580
WT
WT 1Dx2
T581
WT
T590
1Dx5 1Bx7* 1By8 1Dy10 1Dy12
T580
WT
WT
T581
WT
T590
Fig. 1. (A) SDSPAGE of seed protein extracts from transgenic wheat lines and their non-transformed parent (WT). The location of the HMW-GS native to Anza (1Dx2, 1Bx7*, 1By9 and 1Dy12) are indicated. (B) Southern blot analysis of the transgenic lines. Genomic DNA from lines T580, T581 and T590 was digested with EcoRV, ScaI and EcoRV, respectively, and hybridized with probes amplied from plasmids p1Ax1, p1Dx5 and pK-Dy10A. Four, three and two extra hybridized bands, compare to non-transformed parent (WT), are present in DNA of lines T580, T581 and T590, respectively.
polymers (F1), lower molecular mass glutenin polymers (F2), ugliadins (F3), a- and g-type gliadins (F4) and albumins and globulins (F5). The sonication procedure results in shearing of glutenin polymers and hence the sizes of the polymers separated by SEHPLC do not accurately reect the sizes of those present in vivo. However, the relative amounts of the peaks do relate to dough strength with %F1/%F2 and (%F3 %F4)/%F1 showing particularly strong correlations (Millar, 2003). All three transgenic lines had higher values for %F1 and %F1/%F2, indicating that they had higher proportions of high molecular mass glutenin polymers. This effect was also greater in line T590 expressing the 1Dy10 subunit than in lines T580 and T581. However, values for (%F3 %F4)/%F1 decreased in all the transgenic lines relative to the untransformed parent (Table 1). 3.2. Technological properties of our The potential breadmaking quality of the ours was determined over 3 years using a range of widely-used grain and dough testing methods (Table 2). The Falling numbers of all of the samples were higher than the recommended limit of 180 s with no differences among all four lines. The grain test weight was also similar for all four lines. The thousand seed weight of line T580 was higher than that of the other lines. The Zeleny sedimentation was increased in all three transgenic lines and this was greatest in the line expressing subunit 1Dy10. The rheological properties and water absorption of the dough were determined using a Brabender Farinograph (Fig. 2) and 10 g Mixograph (Fig. 3) while the dough viscoelastic properties were determined using a Chopin Alveograph (Fig. 4). The Farinograph
characteristics of transgenic lines are shown in Table 2. No signicant differences were detected for water absorption but the rheological properties differed with the arrival and peak times, increasing signicantly in all three transgenic lines in comparison to the control line. The transgenic lines expressing subunits 1Ax1 and 1Dx5 gave similar values for arrival time and peak time while the line expressing subunit 1Dy10 gave the highest values for these two parameters (Table 2). The dough stability and departure time were also signicantly higher for the line transformed with subunit 1Dy10. Dough stability is an indicator of the overall quality of the protein in the our and this increased from 1.3 in the control line to 11.9 in the line expressing subunit 1Dy10 (Table 2). The mixing tolerance index (MTI) is an indicator of how well the dough will perform during the critical nal stages of mixing and is related to strength (Shuey, 1984). Thus, a high MTI means that the dough will tend to break down quickly while a low MTI may indicate that the dough will require longer mixing to fully develop. The MTI was decreased signicantly in all three transgenic lines with respect to that of the control line. However, this decrease was greater for transgenic line T590 expressing subunit 1Dy10, for which the MTI value fell from 132.8 (BU) in the control line to 35.2 (BU) (Table 2). The degree of softening varied from 132.5 (BU) in the control line to 56.1 (BU) in the line transformed with subunit 1Dy10. Higher values of degree of softening, measured in Brabender units, are characteristic of weak ours and lower values of strong ours. The rheological properties of the transgenic wheats were also determined using a 10 g Mixograph (Fig. 3). The Mixograph is widely used in cereal research as it measures a range of rheological parameters that relate to the behaviour of the dough in bread
152
Table 1 Characterization of the storage proteins in transgenic and control lines of wheat cultivar Anza Parameter Line Anza Endogenous HMW Transgenic HMW Number of insertions Crude protein (%)a Glutenin (mg mg1 our)b Gliadin (mg mg1 our)b Glutenin/gliadin HMW compositionc 1Ax1 1Dx2 1Dx5 1Bx7* 1By8 1Dy10 1Dy12 HMW (% glutenins) LMW (% glutenins) HMW/LMW ratio 1Dx2/1Dy12 ratio 1Bx7/1By8 ratio x-type/y-type ratio SE-HPLCd %F1 %F1/%F2 (%F3 %F4)/%F1 7* 8, 2 12 NA NA 11.7 0.3 4.2 0.2 2.7 0.1 1.53 0.04 NA 27.8 0.6 NA 27.7 0.9 19.4 0.9 NA 25.1 0.4 31.7 0.79 68.3 0.61 0.47 0.01 1.11 0.03 1.43 0.11 1.25 0.06 T580 7* 8, 2 12 1Ax1 4 13.2 0.4 3.3 0.2 3.4 0.1 0.96 0.05 25.2 0.6 16.8 0.5 NA 22.2 0.3 15.5 0.3 NA 20.3 0.8 32.4 1.21 67.6 1.13 0.49 0.02 0.83 0.05 1.43 0.04 1.80 0.07 T581 7* 8, 2 12 1Dx5 2 13.7 0.6 3.6 0.4 3.7 0.4 0.98 0.06 NA 19.1 0.4 20.3 0.8 23.4 0.5 15.6 0.9 NA 21.6 0.9 36.8 0.52 63.2 0.49 0.58 0.01 0.88 0.06 1.50 0.12 1.69 0.06 T590 7* 8, 2 12 1Dy10 3 12.2 0.3 5.3 0.2 3.5 0.1 1.51 0.04 NA 21.5 0.7 NA 26.2 0.2 15.9 0.9 17.9 0.6 18.5 0.7 32.9 0.45 67.1 0.51 0.49 0.01 1.16 0.07 1.64 0.09 0.91 0.03
12.7 0.35 0.60 0.02 4.08 0.12
13.6 0.35 0.61 0.01 3.63 0.12
13.0 0.05 0.62 0.00 3.88 0.03
14.1 0.30 0.68 0.01 3.54 0.10
NA, not applicable. a Kjeldahl method. Average of six replications standard error. b Glutenin and gliadins were determined using a Bradford assay system. Average of 30 replications standard error. c HMW composition was determined by densitometry analysis. Average of 30 replications standard error. d SE-HPLC was carried out in duplicate by bulking samples from the third year. Average standard error.
making and other food processing systems. The machine measures the resistance as the dough is mixed, the most important parameters being the mixing time (time to maximum resistance), the peak resistance (maximum resistance obtained during mixing) and the resistance breakdown (the loss of resistance on over-mixing). In
Table 2 Technological properties of the transgenic and control lines of wheat cultivar Anza Parametera Line Anza Falling number (s) Grain test weight (g l1) Thousand seed weight (g) Zeleny sedimentation (ml) Farinograph Water absorption Arrival time (min) Peak time (min) Dough stability (min) Departure time (min) Mixing Tolerance Index (BU) Degree of softening (BU) Mixograph Mixing time (s) Peak resistance (AU) Resistance breakdown (%) Alveograph Tenacity, P (mm) Extensibility, L (mm) P/L ratio W-value (104 J)
a
general, strong doughs have long mixing times, high peak resistances and low resistance breakdown. Transformation with the three HMW subunit genes led in all cases to increased dough strength, with quantitative differences depending on the subunit expressed (Fig. 3, Table 2). Thus, the mixing time increased from
T580 350 9.8 79.8 5.0 42.1 1.6 28.2 2.0
T581 350 3.0 79.4 4.3 37.5 0.7 31.3 3.8
T590 346 1.2 80.2 4.1 39.7 0.9 41.1 3.0
350 9.5 79.9 3.0 38.4 0.9 20.2 0.4
72.2 1.5 2.1 0.1 2.7 0.3 1.3 0.3 3.5 0.4 132.8 14.4 132.5 18.9
72.8 0.9 3.4 0.6 4.4 0.5 2.7 0.2 6.0 0.6 77.5 9.2 76.7 8.5
73.8 2.8 3.7 0.2 4.8 0.4 2.7 0.3 6.3 0.6 80.8 7.6 71.7 9.2
73.7 1.2 4.6 0.3 6.6 0.8 11.9 2.2 16.5 2.1 35.2 3.8 56.1 6.2
30.0 58.0 37.9
72.0 75.0 26.7
87.0 60.0 18.3
120.0 61.5 18.7
42.2 5.1 94.2 4.5 0.4 0.03 97 8.2
65.8 3.1 91.8 3.0 0.7 0.04 170 13.2
71.7 3.7 61.0 0.6 1.2 0.07 180 12.1
79.0 5.4 82.3 5.9 1.0 0.12 258 4.9
Values are the average of six replications standard error, except for Mixograph parameters for which bulked samples from the third year were used.
153
Fig. 2. Farinograms of the doughs prepared from non-transformed control (Anza) and the our from transgenic lines T580, T581 and T590, expressing the HMW-GS 1Ax1, 1Dx5 and 1Dy10, respectively.
30 s in the control line to 72 in that with subunit 1Ax1, to 87 in that with subunit 1Dx5, and to 120 in that with subunit 1Dy10 (Table 2). However, the peak resistance values were similar for all lines except for the line T580 expressing subunit 1Ax1 where it was higher (Table 2). The resistance breakdown decreased from 37.9% in the control line to 26.7, to 18.3, and to 18.7 in the lines with subunits 1Ax1, 1Dx5 and 1Dy10, respectively (Table 2). The Chopin Alveograph is a dough-testing instrument that simulates the expansion of bubbles in fermenting dough and as such the parameters derived from this test should give some indication as to how the dough will act during fermentation. The
most important parameters obtained from an alveogram are the tenacity (P) of the dough (expressing the resistance of the dough to deformation), the dough extensibility (L), the P/L ratio (which gives a general indication of the viscoeslastic properties of the dough) and the deformation energy (W), which is proportional to the energy required for deformation. This W-value is used to measure the dough strength by bakers. In general, low W values are related to weak ours and high values to strong ours. Transformation with the HMW subunit genes led to increased tenacity (P) in all three transgenic lines with respect to the control line (Table 2). However, the dough extensibility (L) decreased signicantly from
Anza
100 100
T580 (1Ax1)
Resistance (AU)
Resistance (AU)
50
50
0 0 90 180 270 360
0 0 90 180 270 360
Time (s) T581 (1Dx5)

100 100
Time (s) T590 (1Dy10)
Resistance (AU)
50
Resistance (AU)
0 90 180 270 360
50
0 0 90 180 270 360
Time (s)
Time (s)
Fig. 3. Mixograph curves of the doughs prepared from non-transformed control (Anza) and the our from transgenic lines T580, T581 and T590, expressing the HMW-GS 1Ax1, 1Dx5 and 1Dy10, respectively. Resistance in arbitrary units (AU) is plotted against time in minutes.
154
Fig. 4. Alveograph curves of the doughs prepared from non-transformed control (Anza) and the our from transgenic lines T580, T581 and T590, expressing the HMW-GS 1Ax1, 1Dx5 and 1Dy10, respectively.
94.2 in the control line to 61.0 in the line transformed with subunit 1Dx5. The P/L ratio increased signicantly from 0.4 in the control line to 0.7, 1.2, and 1.0 in that for lines transformed, respectively, with subunits 1Ax1, 1Dx5 and 1Dy10 (Table 2). Finally, the W value increased from 96.7 in the control line to 170.0, 180.3, and 258.3 in the lines with subunits 1Ax1, 1Dx5 and 1Dy10, respectively. 4. Discussion Dough mixing and development are critical processes in bread production. Much research has therefore been conducted to determine the parameters that inuence them and to facilitate the production of high quality bread that satises not only processing requirements but also customer expectations. The HMW-GS of wheat are major determinants of the breadmaking quality of wheat and, in particular, the subunit pair 1Dx5 1Dy10 and subunit 1Ax1 have high quality scores (Payne, 1987; Shewry et al., 2003). The production of transgenic lines of wheat differing in their HMW-GS composition makes it possible to determine the effects of individual subunits on the rheological and technological properties of wheat gluten. However, the behaviour of individual subunits may also be affected by the HMW-GS composition in which they are expressed as subunit interactions are crucial in forming the glutenin polymers. In this work we report the effects of the HMW-GS 1Ax1, 1Dx5 and 1Dy10 on the breadmaking quality of the bread wheat cultivar Anza that contain the HMW-GS pairs 1Dx2 1Dy12 and 1Bx7* 1By8 and is null for the Glu-A1 locus. Wheat cultivars containing subunits 1Dx2 1Dy12 generally have lower dough resistance than those containing subunits 1Dx5 1Dy10 (Shewry et al., 2003). Therefore, this genotype is an excellent background to determine separately the effects of subunits 1Dx5 and 1Dy10 and subunit 1Ax1 on breadmaking quality. Although subunits 1Dx5 and 1Dy10 have been separately introduced into the bread wheat Bobwhite (Blechl et al., 2007), this cultivar already contain both subunits which could inuence the behaviour of the additional subunit protein encoded by the transgenes. This is therefore the rst report on the characterization of wheat expressing subunit 1Dy10 in the absence of subunit 1Dx5. The
results show that subunit 1Dy10 had a greater effect on breadmaking quality than subunits 1Ax1 and 1Dx5. However, the expression of subunits 1Ax1 and 1Dx5 also led to increased breadmaking quality. This is not surprising for subunit 1Ax1 as this protein has been reported to have positive effects in a range of wheat backgrounds (Barro et al., 2003; Darlington et al., 2003). However, the benecial effect of subunit 1Dx5 was unexpected as previous transgenic wheat lines expressing this subunit had overstrong doughs unsuitable for breadmaking (Darlington et al., 2003; Rooke et al., 1999b). The expression levels of the transgenic 1Ax1, 1Dx5 and 1Dy10 HMW-GS were comparable to those of the endogenous subunits. However, although lines T581 and T590 expressed similar levels of subunits 1Dx5 and 1Dy10 (Table 1) they had different effects on other aspects of protein composition, including changes in the ratios of glutenin/gliadin, the HMW/LMW, the 1Dx2/1Dy12, x-type/ y-type HMW-GS and the proportions of high molecular mass glutenin polymers (Table 1). In contrast, the lines transformed to express subunits 1Ax1 and 1Dx5 showed more similar changes in protein composition, with higher protein contents and decreased ratios of glutenin/gliadin and of 1Dx2/1Dy12. In addition, both transgenic lines showed similar increases in the ratio of x-type/ytype subunits compared to the control line (Table 1) while the line expressing subunit 1Dx5 also showed a signicant increase in the HMW to LMW ratio. The increases in the HMW/LMW ratio and the changes in the glutenin/gliadin ratio as a consequence of the expression of transgenic HMW-GS have been reported previously (Alvarez et al., 2001; Blechl et al., 2007; Rakszegi et al., 2005). These changes also indicate that the new HMW-GS affect the structure of the glutenin network with the x-type subunits (i.e. 1Ax1 and 1Dx5) having similar effects which are different to those of subunit 1Dy10. In the case of T590 (expressing 1Dy10) the composition is more comparable to that of the control line. For example, both the control line and line T590 had similar protein contents and ratios of glutenin/gliadin, HMW/LMW and 1Dx2/1Dy12 (Table 1). However, the x-type/y-type ratio was lower in line T590 than in the control line while the proportion of high molecular mass polymers was signicantly increased.
155
The transgenic lines were analysed over 3 years using standard dough testing methods (Farinograph, Mixograph and Alveograph). This conrmed that the expression of all three subunits increased the dough strength (and hence breadmaking quality) of the cultivar Anza. Furthermore, a stepwise increase was observed depending on the HMW-GS expressed. Subunits 1Ax1 and 1Dx5 resulted in similar changes in the protein composition and structure, as described above, and this is reected in their similar effects on breadmaking quality. Thus, the Farinograph showed increased dough resistance and stability and increased dough tolerance to mixing, the values being characteristic of medium grade ours (Table 2). Improvements in the rheological properties of the transgenic lines expressing subunits 1Ax1 and 1Dx5 were conrmed using the 10 g Mixograph (Fig. 3). Both transgenic lines had longer mixing times and lower values for resistance breakdown than the control line (Table 2). Subunit 1Ax1 has previously been reported to either increase dough resistance (Barro et al., 2003; Darlington et al., 2003) or to have little or no effect on dough resistance (Alvarez et al., 2001). In the former case, the genotype was derived from an Olympic Gabo cross and contained no 1Ax subunit and in the latter report the genotypes (BW and Federal) both expressed subunit 1Ax2*. However, the expression of subunit 1Ax2* was suppressed in the transgenic line of Federal and the breadmaking quality was comparable to those of control lines expressing subunit 1Ax2*. It was therefore suggested that subunit 1Ax1 had a similar effect to subunit 1Ax2*. Most previous studies of lines expressing the 1Dx5 transgene reported overstrong dough properties which were unsuitable for breadmaking. In contrast, in the present study the expression of subunit 1Dx5 in the breadmaking variety Anza resulted in improved dough properties for breadmaking. Rooke et al. (1999b) suggested that although the transgenic subunits are incorporated into glutenin polymers, their organization within the polymers may differ from that in the corresponding control lines. An imbalance in the ratio of x-type/y-type subunits may therefore be responsible for the effects of expression of the 1Dx5 transgene without expression of subunit 1Dy10 (Alvarez et al., 2001). Our results indicate that subunits 1Dx5 and 1 Ax1 are similar in this respect, having similar effects on protein composition and dough properties. In contrast, transformation with subunit 1Dy10 led to stronger ours than those made from the control line and lines transformed with subunits 1Ax1 and 1Dx5. Farinograph parameters such as dough stability and peak time were increased by 9.2-fold and 2.4fold, respectively, in line T590 (expressing 1Dy10) with respect to the control line, while the mixing tolerance index and the degree of softening decreased by 3.8-fold and 2.4-fold, respectively. The Mixograph and Alveograph parameters similarly showed that subunit 1Dy10 had a greater effect on breadmaking quality. Thus, the mixing time was increased by four-fold and resistance breakdown decreased by two-fold in line T590 compared with the control line. The W value was also increased by 2.7-fold in line T590 compared to the control line. In this study, subunits 1Dx5 and 1Dy10 were compared in a cultivar of low breadmaking quality which provides an appropriate background to explore their effects on dough properties. Small scale mixing experiments incorporating puried 1Dx5 and 1Dy10 subunits into dough (Uthayakumaran et al., 2000) showed that subunit 1Dx5 resulted in longer mixing times and lower resistance breakdown than subunit 1Dy10. In contrast, our results showed that expression of subunits 1Dx5 and 1Dy10 in transgenic wheat resulted in doughs with similar low resistance breakdown but that dough made from our expressing the subunit 1Dy10 transgene had a longer mixing time. This effect of the subunit 1Dy10 transgene on dough strength is in agreement with the results reported by Blechl et al. (2007) for the expression of subunit 1Dy10 in Bobwhite. This expression of subunit 1Dy10 was also
associated with a greater effect on the proportion of high molecular mass glutenin polymers than those observed with subunits 1Dx5. Acknowledgements The authors acknowledge funding by the Spanish CICYT (project AGL2007-65685-C02-01). Rothamsted Research receives grantaided support from the Biotechnology and Biological Sciences Research Council (BBSRC) of the UK. Plasmid pK-Dy10A was provided by Dr. Ann E. Blechl. The technical assistance of Azahara Vida is also acknowledged. References
AACC, 2000. Physical Dough Tests. Mixograph Method tenth ed Method 54-40A. Altpeter, F., Vasil, V., Srivastava, V., Vasil, I.K., 1996. Integration and expression of the high molecular weight glutenin subunit 1Ax1 gene into wheat. Nature Biotechnology 14, 11551159. Altpeter, F., Popelka, J.C., Wieser, H., 2004. Stable expression of 1Dx5 and 1Dy10 high-molecular-weight glutenin subunit genes in transgenic rye drastically increases the polymeric glutelin fraction in rye our. Plant Molecular Biology 54, 783792. Alvarez, M.L., Gomez, M., Carrillo, J.M., Vallejos, R.H., 2001. Analysis of dough functionality of ours from transgenic wheat. Molecular Breeding 8, 103108. Anderson, O.D., Greene, F.C., Yip, R.E., Halford, N.G., Shewry, P.R., MalpicaRomero, J.M., 1989. Nucleotide sequences of the two high-molecular-weight glutenin genes from the D-genome of a hexaploid bread wheat, Triticum aestivum L. cv Cheyenne. Nucleic Acids Research 17, 461462. Barcelo, P., Lazzeri, P.A., 1995. Transformation of cereals by microprojectile bombardment of immature inorescence and scutellum tissues. In: Jones, H. (Ed.), Methods in Molecular Biology: Plant Gene Transfer and Expression Protocols. Humana Press, Totowa, NJ, pp. 113123. Barro, F., Rooke, L., Bekes, F., Gras, P., Tatham, A.S., Fido, R., Lazzeri, P.A., Shewry, P.R., Barcelo, P., 1997. Transformation of wheat with high molecular weight subunit genes results in improved functional properties. Nature Biotechnology 15, 12951299. Barro, F., Cannell, M.E., Lazzeri, P.A., Barcelo, P., 1998. The inuence of auxins on transformation of wheat and tritordeum and analysis of transgene integration patterns in transformants. Theoretical and Applied Genetics 97, 684695. Barro, F., Barcelo, P., Lazzeri, P.A., Shewry, P.R., Martn, A., Ballesteros, J., 2002. Field evaluation and agronomic performance of transgenic wheat. Theoretical and Applied Genetics 105, 980984. Barro, F., Barcelo, P., Lazzeri, P.A., Shewry, P.R., Ballesteros, J., Martin, A., 2003. Functional properties of ours from eld grown transgenic wheat lines expressing the HMW subunit 1Ax1 and 1Dx5 genes. Molecular Breeding 12, 223229. Blechl, A.E., Anderson, O.D., 1996. Expression of a novel high-molecular-weight glutenin subunit gene in transgenic wheat. Nature Biotechnology 14, 875879. Blechl, A., Lin, J., Nguyen, S., Chan, R., Anderson, O.D., Dupont, F.M., 2007. Transgenic wheats with elevated levels of Dx5 and/or Dy10 high-molecular-weight glutenin subunits yield doughs with increased mixing strength and tolerance. Journal of Cereal Science 45, 172183. Bradford, M.M., 1976. A rapid and sensitive method for quantitation of microgram quantities of protein utilizing the principle of protein-dye-binding. Analytical Biochemistry 72, 248254. Christensen, A.H., Quail, P.H., 1996. Ubiquitin promoter-based vectors for high-level expression of selectable and/or screenable marker genes in monocotyledonous plants. Transgenic Research 5, 213218. Darlington, H., Fido, R., Tatham, A.S., Jones, H., Salmon, S.E., Shewry, P.R., 2003. Milling and baking properties of eld grown wheat expressing HMW subunit transgenes. Journal of Cereal Science 38, 301306. DOvidio, R., Anderson, O.D., 1994. PCR analysis to distinguish between alleles of a member of a multigene family correlated with wheat bread-making quality. Theoretical and Applied Genetics 88, 759763. Gupta, R.B., MacRitchie, F., 1994. Allelic variation at glutenin subunit and gliadin loci, Glu-1, Glu-3 and Gli-1 of bread wheats: biochemical basis of the allelic effects on dough properties. Journal of Cereal Science 19, 1928. Halford, N.G., Field, J.M., Blair, H., Urwin, P., Moore, K., Robert, L., Thompson, R., Flavell, R.B., Tatham, A.S., Shewry, P.R., 1992. Analysis of HMW glutenin subunits encoded by chromosome 1A of bread wheat (Triticum aestivum L.) indicates quantitative effects on grain quality. Theoretical and Applied Genetics 83, 373378. ICC, 1992a. Method for using of the Chopin-Alveograph. International Association for Cereal Science and Technology Method No. 121. ICC, 1992b. Method for using the Brabender Farinograph. International Association for Cereal Science and Technology Method No. 115/1. ICC, 1994. Determination of the sedimentation value (according to Zeleny) as an approximate measure of baking quality. International Association for Cereal Science and Technology Method No. 116/1. ICC, 1995. Determination of the Falling Number according to Hagberg-Perten as a mesure of the degree of Alpha-Amylase activity in grain and our. International Association for Cereal Science and Technology Method No. 107/1.
156
E. Leon et al. / Journal of Cereal Science 49 (2009) 148156 tritordeum by transformation with HMW glutenin subunit genes. Theoretical and Applied Genetics 99, 851858. Rooke, L., Bekes, F., Fido, R., Barro, F., Gras, P., Tatham, A.S., Barcelo, P., Lazzeri, P., Shewry, P.R., 1999b. Overexpression of a gluten protein in transgenic wheat results in greatly increased dough strength. Journal of Cereal Science 30, 115120. Sangtong, V., Moran, D., Chikwamba, R., Wang, K., Woodman-Clikeman, W., Long, M., Lee, M., Scott, M., 2002. Expression and inheritance of the wheat Glu1DX5 gene in transgenic maize. Theoretical and Applied Genetics 105, 937945. Shewry, P.R., Tatham, A.S., Fido, R.J., 1995. Separation of plant proteins by electrophoresis. In: Jones, H. (Ed.), Methods in Molecular Biology: Plant Gene Transfer and Expression Protocols. Humana Press, Totowa, NJ, pp. 399422. Shewry, P.R., Halford, N.G., Tatham, A.S., Popineau, Y., Laandra, D., Belton, P.S., 2003. The high molecular weight subunits of wheat glutenin and their role in determining wheat processing properties. Advances in Food and Nutrition Research 45, 219302. Shuey, W.C., 1984. Interpretation of the Farinogram. In: DAppolonia, B.L., Kunerth, W.H. (Eds.), The Farinograph Handbook. American Association of Cereal Chemistry, St. Paul, MN, pp. 3132. Stacey, J., Isaac, P.G., 1994. Isolation of DNA from plants. In: Isaac, P.G. (Ed.), Methods in Molecular Biology: Protocols for nucleic acid analysis by nonradiactive probes. Humana Press, Totowa, NJ, pp. 915. Uthayakumaran, S., Stoddard, F.L., Gras, P.W., Bekes, F., 2000. Effects of incorporated glutenins on functional properties of wheat dough. Cereal Chemistry 77, 737743.
Millar, S.J., 2003. The development of near infrared (NIR) spectroscopy calibrations for the prediction of wheat and our quality. Project Report No. 310. HGCA, UK. 101. Morel, M.-H., Dehlon, P., Autran, J.C., Leygue, J.P., Bar-LHelgouach, C., 2000. Effects of temperature, sonication time and power settings on size distribution and extractability of total wheat our proteins as determined by size-exclusion high-performance liquid chromatography. Cereal Chemistry 77, 685691. Murashige, T., Skoog, F., 1962. A revised medium for rapid growth and bio assays with tobacco tissue cultures. Physiologia Plantarum 15, 473497. Payne, P.I., 1987. Genetics of wheat storage proteins and the effect of allelic variation on bread-making quality. Annual Review of Plant Physiology 38, 141153. Popineau, Y., Cornec, M., Lefebvre, J., Marchylo, B., 1994. Inuence of high Mr glutenin subunits on glutenin polymers and rheological properties of glutens and gluten subfractions of near-isogenic lines of wheat Sicco. Journal of Cereal Science 19, 231241. Popineau, Y., Deshayes, G., Lefebvre, J., Fido, R., Tatham, A.S., Shewry, P.R., 2001. Prolamin aggregation, gluten viscoelasticity, and mixing properties of transgenic wheat lines expressing 1Ax and 1Dx high molecular weight glutenin subunit transgenes. Journal of Agricultural and Food Chemistry 49, 395401. Rakszegi, M., Bekes, F., Lang, L., Tamas, L., Shewry, P.R., Bedo, Z., 2005. Technological quality of transgenic wheat expressing an increased amount of a HMW glutenin subunit. Journal of Cereal Science 42, 1523. Rooke, L., Barro, F., Tatham, A.S., Fido, R., Steele, S., Bekes, F., Gras, P., Martin, A., Lazzeri, P.A., Shewry, P.R., Barcelo, P., 1999a. Altered functional properties of

Wheat glutenin proteins assemble into a nanostructure with unusual structural features
Sarah H. Mackintosh a, Susie J. Meade a, Jackie P. Healy b, Kevin H. Sutton a, Nigel G. Larsen a, Adam M. Squires c,1, Juliet A. Gerrard b, *
a b c
Food and Biomaterials Innovation, New Zealand Institute for Crop & Food Research Ltd., Private Bag 4704, Christchurch, New Zealand School of Biological Sciences, University of Canterbury, 20 Kirkwood Ave, Private Bag 4800, Christchurch 8020, New Zealand Cavendish Laboratory, Cambridge University, Cambridge, UK
Article history: Received 11 December 2007 Received in revised form 19 August 2008 Accepted 22 August 2008 Keywords: Glutenins X-ray bre diffraction Wheat protein Protein nanostructure
a b s t r a c t
The proteins of wheat have a known propensity to aggregate into a variety of forms. We report here a novel nanostructure from wheat proteins, derived from a crude extract of high molecular weight glutenins. The structure was characterised by a signicant thioavin T (ThT) uorescence and a brillar morphology by transmission electron microscopy (TEM). The ThT uorescence and TEM data are suggestive of an amyloid structure, but the X-ray bre diffraction data show a reection pattern (4.02, 4.24.3, 4.6, 12.9, 19.3 and 38.7 ) inconsistent with both the classic amyloid form and the previously described b-helix structure. The 4.6 reection is consistent with that predicted for the amyloid inter-bstrand, and the absence of the inter-b-sheet distance at z1011 is not unprecedented in amyloid-like structures. However, our observed X-ray reection pattern has not been previously reported and suggests a novel wheat glutenin nanostructure. 2008 Elsevier Ltd. All rights reserved.
1. Introduction Proteins have an innate ability to adopt specic three dimensional forms with a plethora of functions. This affords the potential for design of nanoscale materials via the manufacture of selfassembled protein nanostructures, with properties that are not easily duplicated with traditional organic molecules and other polymers (Sanford and Kumar, 2005). Signicantly, protein nanomaterials incorporate surface chemical functionality via the amino acid side chains, therefore providing additional possibilities for the construction of advanced nanomaterials, with a combination of structural and (bio)chemical function. The enormous promise for the use of peptides and protein as smart materials has attracted much recent attention (Shen et al., 2006; Waterhouse and Gerrard, 2004). Amyloid brils are highly ordered nanobres of insoluble protein, rich in b-strand content (Dobson, 1999). The mature bril comprises a number of protolaments (sub-structures) that typically twist together to form a nanostructure with a diameter of
* Corresponding author. Tel.: 64 3 3667001; fax: 64 3 3642590. E-mail address: juliet.gerrard@canterbury.ac.nz (J.A. Gerrard). 1 Present address: School of Chemistry, University of Reading, Whiteknights, Reading RG6 6AD, UK. 0733-5210/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2008.08.003
approximately 515 nm, and a rope-like appearance when imaged at high magnications (Serpell, 2000). Whilst there are a number of competing structural models (Makin and Serpell, 2005), amyloid brils are dened by observations of characteristic electron microscopy, specic chemical staining and X-ray bre diffraction. The specic arrangement of the b-strands within the mature bril is dened as a cross-b pattern and leads to characteristic reections in the X-ray bre diffraction pattern of these brils, which is increasingly accepted as a diagnostic indicator of the presence of amyloid brils (Makin and Serpell, 2004). As well as their well recognised importance in the diseases of protein misfolding (Dobson, 2001), the use of amyloid brils as nanomaterials is attracting increasing attention (Gras, 2007; Waterhouse and Gerrard, 2004). Other amyloid-like nanostructure models have been proposed, including b-helix and spiral structures (DeMarco et al., 2006; Makin and Serpell, 2005). Since polyglutamine proteins are associated with amyloid formation in the body (Chen et al., 2002), our attention was drawn to the proteins of wheat as a potential raw material for amyloid bril manufacture. Wheat proteins are known to have a high glutamine content (Sugiyama et al., 1985), including the high molecular weight glutenin subunits (HMW-GS) associated with our quality which have a (glutamic acid glutamine content) in the order of 30% (Anjum et al., 2008; Kipp et al., 1996). A BLAST search (Altschul et al., 1997) on the wheat (Triticum aestivum)
158
S.H. Mackintosh et al. / Journal of Cereal Science 49 (2009) 157162
genome returned matches for polyglutamine sequences up to eleven amino acids in length. Additionally, gluten proteins are well known to form aggregated structures; indeed it is this ability that facilitates their use in the production of a variety of baked goods and pasta products (Shewry et al., 2002). They have also been explored for biomaterial use (Guilbert et al., 2001). The gluten network itself is extremely complex, with the molecular weight of these HMW-GS components ranging from approx. 6788 kDa (Shewry et al., 2002). The units themselves appear to be comprised of three distinct domains: a central repetitive domain of variable length (440680 residues), and much shorter anking non-repetitive terminal domains (N 81104 and C 42 residues) (Tatham, 1995). Structural information to date has been based on predictive modeling, microscope imaging, indirect methods such as FTIR and CD, and limited SAXS analysis (Lindsay and Skerritt, 1999). Indications are that the secondary structure of the proteins in the matrix includes a high proportion of b-turn structure for the central domain, while the terminal domains comprise a-helices (Tatham, 1995 and references therein). The low molecular weight group of glutenin proteins (LMW-GS) is less well characterised. It is known they have a glutamine rich N-terminal domain (again predicted to form b-reverse turns), and are rich in cysteine residues located primarily in the C-terminal end, but also in the central domain (predicted to be a-helical). This group of proteins predominantly exists in monomeric form, and contains numerous internal disulde bonds (Lew et al., 1992). Subgroups within the LMW-GS are also known to possess similar secondary structure to the gliadin protein group, which has been documented to aggregate under appropriate conditions (Kasarda et al., 1967). The gliadin bril aggregates were observed to be 80 thick, up to several thousand ngstroms long and dissociated to globular protein subunits at very low ionic strength and low pH. Despite considerable research into their aggregation properties, the tendency for gluten proteins to form amyloid-like structures has not been previously explored. We report herein an exploration of whether or not gluten proteins can form amyloid structures and the characterisation of a novel nanostructure derived from wheat glutenins, characterised by spectroscopic binding assays, transmission electron microscopy and X-ray bre diffraction. 2. Experimental 2.1. Materials Unless otherwise stated, all chemicals and reagents were purchased from Sigma Chemical Company Ltd. (St Louis, U.S.A.), Aldrich Chemicals (Milwaukee, U.S.A.) or BDH Laboratory Supplies (Poole. U.K.), and were of analytical grade. The wheat our derived protein was extracted from a Domino cultivar, sourced by Crop and Food Research Ltd. and milled in-house. 2.2. Extraction protocol Wheat proteins were crudely fractionated using an extraction method based on the methods of Gerrard et al. (2001), which were derived from the protocols of Osborne (1907). The albumins and globulins were extracted in dilute saline (2% w/v). The mixture was mixed by pulse vortex every 5 min for 30 min to allow extraction of the appropriate protein group, then centrifuged at 10 000 g for 5 min. The resulting pellet was resuspended in a concentrated ethanol solution (70% v/v) and mixed by pulse vortex every 5 min for 30 min to resuspend the pellet and to separate the soluble and insoluble components. After centrifugation (10 000 g, 5 min) the gliadin extract was decanted. The pellet was resuspended in a SDSphosphate buffer (0.05% w/v SDS, 0.05 M phosphate, pH 6.9) with mixing by pulse vortex (every 5 min for 30 min), to extract the
LMW-GS and HMW-GS. After centrifugation (10 000 g, 5 min) the pellet was again resuspended in SDS-phosphate buffer (0.05% w/v SDS, 0.05 M phosphate, pH 6.9) with mixing by pulse vortex (every 5 min for 30 min). This nal solution of glutenins was sonicated at 30 W for 15 s using a Branston model 250 sonic disruptor, to solubilise any protein components still held within the pellet. 2.3. Protein concentration Protein concentrations were determined using a modied version of the Bradfords (1976) method. A standard curve was determined using bovine serum albumin. 2.4. PAGE Polyacrylamide gel electrophoresis was routinely run using an Invitrogen XCell SureLock Mini-Cell powered by a Bio-Rad 300 power pack, and precast 412% BisTris gels from Invitrogen. The gel system was run in a 1 3-(N-morpholino) propanesulfonic acid (MOPS) SDS running buffer, over approx. 90 min at 150 V, and stained with Coomassie blue. The marker was wide molecular range SigmaMarker (product no. S8445 from Sigma). 2.5. Screening protocols In preliminary screens, all four fractions of wheat protein were incubated in an array of solution conditions and periodically tested using thioavin T (ThT, see below) for the presence of amyloid brils. In particular, additions of the following compounds to the extraction buffers were tested: acid (HCl or H2SO4); sodium chloride (00.2 M); denaturing compounds (26 M urea, 030% (v/v) TFE, 1 M mercaptoethanol or 0.5% (w/v) SDS); and insulin brils (2% v/v 5.8 mg/ml solution of preformed insulin brils/wheat protein solution). The extracted glutenin protein fractions proved to be the most interesting and were explored further. No further experiments were carried out on the albumin and globulin or gliadin fractions. Glutenin fractions were then extracted, and the proteins lyophilised and resuspended at a concentration of 10 mg/ml (w/v) in a variety of modied buffers. A screen of conditions was trialled in order to optimise conditions for bril formation. The solution conditions of the glutenin extracts were varied in order to optimise the conditions for bril formation, as judged by the ThT assay. A range of conditions and variables were investigated both individually and in combination and they are summarized in Table 1. The solutions were then incubated at temperatures of either 25 C or 37 C and monitored for periods of up to 105 days and compared to control samples kept frozen at 20 C. As required, trypsin solution was prepared with 1 mM HCl and mercaptoethanol (10 ml/ml) and the glutenin extracts were treated at a ratio of 1:20 (v/v) prior to incubation. Insulin brils were formed by incubation of insulin from bovine pancreas (Sigma product no.
Table 1 Range of conditions for glutenin incubations. Variable Glutenin concentration Temperature Time pH Urea TFE Sodium chloride Sonication Trypsin pre-treatment Insulin seeding Range 210 mg/ml 20 to 37 C 0105 days 27 06 M 030% 00.2 M / / 2% v/v (5.8 mg/ml solution of preformed insulin brils/wheat protein solution)
159
I5500) at pH 2 (HCl in distilled water) at 50 C for 48 h according to the method of Nilsson and Dobson (2003). Fibril formation was conrmed by TEM and ThT assay. These seed brils were added to selected glutenin samples prior to incubation. In addition to changing the solution conditions, the impact of shaking and/or periodic sonication during the incubation was explored. Aliquots removed from the incubating reactions were stored at 20 C until required for analysis by ThT and TEM (below). 2.6. ThT assay The ThT method used was based on the protocols of LeVine (1999). The protein samples were diluted to a concentration of 10 20 mg/ml in Tris buffer (50 mM Tris, 100 mM NaCl, pH 7.5). To this solution, the ThT dye was added to a nal concentration of 5 mM. The solution was mixed and left to stand at room temperature for 3 min to allow binding between the dye and protein to equilibrate, before the emission spectra were recorded. The spectra of buffer only, buffer with ThT and buffer with protein were used as the experimental controls. Fluorescence spectroscopy was performed using a Cary Eclipse Varian uorescence spectrophotometer interfaced with Cary Eclipse operating software (V 2.0), using 1 cm path length quartz cuvettes and the excitation was set to 450 10 nm and emission set to 470520 nm. The relative uorescence is graphed. This reading is the incubated sample uorescence with the control (non-incubated protein with ThT) uorescence subtracted. The error bars are standard errors for a triplicate measurement. 2.7. Transmission electron microscopy Samples were analysed by TEM, based on the protocols described by Brenner and Horne (1959), to conrm the presence of brils. The preparation involved placing small aliquots (5 ml) of protein bril samples on formvar-coated copper grids. After 90 120 s, the excess solution was drained with lter paper, the grids washed with double distilled water and the samples negatively stained with 1% (w/v) uranyl acetate for 4560 s, before the excess solution was again drained with lter paper and the grids left to dry. For each sample, four images were recorded and a representative image was selected from amongst these. 2.8. X-ray diffraction The glutenin samples were prepared for X-ray bre diffraction by placing a 10 ml droplet of the incubated protein solution between two wax-lled glass capillary ends (Serpell et al., 1999). The capillaries were manually separated as the sample dried, to promote bril alignment. The resulting bre was then analysed using an Xray beam produced by a Rigaku Cu-K rotating-anode source with the data collected using an R-AXIS IV image-plate X-ray detector. The data were analysed using Fit2D (V12.043) software. Images were recorded in the Biochemistry Department, University of Cambridge. 3. Results and discussion 3.1. Screening for bril-forming conditions All four wheat protein fractions the albumins and globulins, the gliadins, the SDS-soluble glutenins and the SDS-insoluble glutenins were put through an initial screen including a range of solution conditions that have been shown to induce bril formation in other systems (Kim et al., 2004). The initial screen showed promising results for the SDS-soluble glutenins and the SDS-insoluble glutenins only (data not shown) and no further
work was carried out on the albumin and globulin and gliadin fractions. The solution conditions of the two glutenin extracts were then varied in a systematic, combinatorial fashion, as described in Section 2. These included acidic conditions and addition of a denaturant. Also explored was the effect of shaking the solution during the incubation, sonicating the solution periodicially during the incubation, and temperature. At various time points throughout the incubation, aliquots were extracted from each of the incubated protein mixture samples and analysed by ThT assay. Evidence for increased b-sheet, consistent with bril-like structures, was found under some conditions, particularly in acidic conditions in the presence of urea, consistent with the requirement for partial unfolding of the proteins prior to bril formation (Chiti and Dobson, 2006). The most successful conditions involved a pH of 67 in the presence of either 2 M urea or 30% (v/v) triuoroethanol. The SDSinsoluble glutenins gave more encouraging results under these conditions than the SDS-soluble glutenins. The SDS-soluble glutenin conditions that gave signicant increases in ThT uorescence are shown in Table 2. An example of the results is shown in Fig. 1. A signicant lag period was observed for all treatments in which positive results were obtained, characteristic of a period of equilibrium where bril growth is at a minimum and interactions between the glutenin subunits are limited and reversible, prior to the growth phase (Nilsson and Dobson, 2003). The lag phase observed is signicantly longer than any lag phase reported for amyloid formation in vitro in the literature to date. This increase in ThT uorescence is consistent with a structure with increased b-sheet content. 3.2. Characterising the aggregated structures: SDS-PAGE In an attempt to establish which of the many proteins in the glutenin fractions were forming the aggregated structure, the incubated solutions were analysed by SDS-PAGE. The solution was spun down to remove any aggregated protein and the supernatant solution analysed as a function of time, for many of the conditions tested. The electrophoretic prole of this fraction is well studied (Gerrard et al., 2001) but typically, no specic proteins were found to be selectively lost from solution; rather, all proteins were lost at a similar rate, although larger proteins were removed from the supernatant somewhat faster than smaller ones. A typical result is found in Fig. 2. These results suggest that a mixture of proteins is incorporated into the aggregate, or that different proteins form different structures that are incorporated into the assembly, and/or that those proteins not incorporated were subject to hydrolysis and/or proteolysis during the long timeframe of the incubation. Extensive further work would be required to distinguish between these possibilities, and due to the long timeframe and low yield of the process, these experiments were not continued further.
Table 2 Summary of SDS-insoluble glutenin incubation conditions that resulted in signicant increase ThT uorescence after 105 days. Frozen controls showed no increase in ThT uorescence. Treatment Temp. ( C) 25 25 37 37 37 37 Conc. (mg/ml) 10 10 10 10 10 10 Conditions pH pH pH pH pH pH 67 67, 67 67, 67, 67, 0.5 M H2SO4 0.5 M HCl 2 M urea 30% (v/v) TFE SDS-insoluble glutenin ThT ThT ThT ThT ThT ThT
160
12 10 8 6 4 2 0 0 10 20 30 40 50 60 70
Time (Days)
Fig. 1. Formation of amyloid-like structures in the SDS-soluble glutenins of wheat incubated at 25 C in the presence of 2 M urea.
3.3. Characterising the aggregated structures: TEM and X-ray bre diffraction ThT is commonly used to monitor amyloid bril formation, as it forms a uorescent species when bound to protein material with a high content of b-structure (LeVine, 1999). A positive ThT result is thus strongly suggestive of the presence of b-sheet structure such as an amyloid bril, but other methods are required to corroborate this observation. Samples which showed a positive ThT test were observed under TEM. A variety of morphologies were apparent in the sample, as shown in Fig. 3. Fibril structures were typically unbranched and of indeterminate length (at least several microns), with widths of between 5 and 15 nm, of similar dimensions to amyloid brils described in the literature (Serpell, 2000). Some displayed the
classical twisted rope appearance (Fig. 3A, B). Consistent with the low yield of ThT positive material (Fig. 1), the nanobrillar structures were found to be contaminated with amorphously aggregated protein. The sticky nature of the amorphous aggregates made it impracticable to separate and purify the nanostructures. We estimate that the yield of these nanostructures was less than 1%. X-ray bre diffraction is increasingly used as a diagnostic indicator for amyloid bril formation (Makin et al., 2005). The diffraction pattern produced when an aligned bril sample is exposed to an X-ray beam can be used to detect the cross-b pattern indicative of b-sheet structure, of indeterminate length, with the b-strands aligned perpendicular to the bril axis and the b-sheets running parallel to it. For amyloid brils, the interstrand separation classically appears as a 4.7 meridonal reection (i.e. in a parallel direction to the mounted bril bre) and the intersheet separation as an equatorial reection at approx. 1011 . However, polyglutamine peptide bre patterns have been reported with a reection at 4.7 , but the 10 reection is often faint or not observed (Perutz et al., 2002). The diffraction pattern obtained from the wheat glutenin protein brillar structure is shown in Fig. 4. It is clear that the sample is anisotropic, but the reections show distinct differences from the classic amyloid form. Equatorially aligned bands at 12.9, 19.3 and 38.7 , and meridionally aligned bands at 4.6, 4.24.3, and 4.02 were observed. A number of fainter bands are also visible. The faint band at 4.6 is suggestive of an aligned structure with an intra-strand distance corresponding to that of a b-sheet. The absence of a diffraction ring at approximately 1011 is also consistent with amyloid brils formed from high glutamine protein. In classic amyloid brils (i.e. those linked with disease) this distance corresponds to the interstrand spacing, and forms an integral part of denition for cross-b-structure. Thus, whilst it is tempting to speculate that the observed nanostructures are amyloid-like, this model does not t all the data. The other wide-angle meridional reections and the inner equatorial reections (12.9, 19.3, and 38.7 ) have not, to our knowledge, been reported in any brillar system; the equatorial reections are more consistent with a lateral assembly (i.e. perpendicular to the bril axis), with a spacing of 38.7 . In light of this, the 4.6 reection may be due to another feature within the brils of a nonamyloid structure, or perhaps indicates that the sample may contain a small amount of amyloid-like brils mixed with nonamyloid brils. The glutenin nanostructure thus appears to contain novel features distinct from both the previously reported classical amyloid bril structure and the hypothesised b-sheet helix structure (DeMarco et al., 2006). 3.4. Attempts to improve the rate and yield of bril formation The novel wheat protein structure was formed at a low rate and the overall yield was too small to be considered as a viable process for commercial manufacturing. All treatments consistently resulted in extremely low yields. In a subsequent screen, therefore, two further methods were explored in order to increase the rate of formation of brillar structures, and boost the yield: seeding the solution with preformed brils, and pre-treatment of the protein with trypsin. An analysis of the sequences of glutenin proteins using the aggregating predicting programme Zygregattor (Pawar et al., 2005) revealed that the regions of local sequence most likely to hinder aggregation of the glutenin proteins were those containing lysine or arginine residues. Since trypsin cuts proteins adjacent to both these residues in a selective manner (Wang and Carpenter, 1967), it was hoped that pre-digestion with trypsin might improve the rate and yield of formation of brils from the wheat protein incubations. Unfortunately, despite conrmation by SDS-PAGE that the trypsin
Fig. 2. SDS-PAGE gel of glutenins, in 2 M urea, incubated for various times up to 70 days. 1. Marker (205, 116, 97.4, 84, 66, 55, 45, 36, 29, 24, 20.1, 14.2, and 6.5 kDa); 2. 7 days; 3. 14 days; 4. 21 days; 5. 28 days; 6. 42 days; 7. 56 days; and 8. 70 days.
Relative Fluorescence
161
Fig. 3. Wheat glutenin protein structure images derived from transmission electron microscopy analysis. Images were taken from SDS-insoluble glutenin proteins incubated at 25 C (2 mg/ml, pH 7.4) for between 54 and 105 days. Scale bars represent 200 nm.
had indeed cut the glutenins into the desired smaller fragments, no subsequent increase in rate of formation of ThT positive species, or yield was observed. Since amyloid formation is a seeded process (Ban et al., 2006), attempts were made to enhance the rate of production by
sonication and addition of preformed bril seeds. These processes have been reported to accelerate bril formation (Stathopulos et al., 2004). The sonication provides strong agitation that can fragment the brils, thus potentially providing more sticky ends for bril elongation. Samples were sonicated periodically during the incubation of wheat protein in a range of different buffer conditions. Insulin seeds were formed by sonication of insulin brils and these insulin seeds were added to the glutenin protein extracts prior to incubation. Neither method was successful. The lack of sequence identity between the insulin seeds and the glutenin structure may account for the failure of the latter method. 3.5. Concluding comments Although our original aim of manufacturing large quantities of protein nanostructures from a readily available heterogeneous protein substrate was not met, a novel aggregated structure has been assembled and characterised from the wheat glutenin proteins, without prior purication. The structure displays some characteristic features of amyloid brils, including a detectable increase in ThT uorescence assays and a brillar morphology when viewed under an electron microscope. While the aggregated material does not display the classic cross-b pattern commonly observed for amyloid material, bands do exist within the pattern at z4.8 (corresponding to inter-b-strand distance). The absence of an inter-b-sheet distance at z1011 is not unprecedented and a number of recent articles describing amyloid-like structures, that may represent a distinct but related form of amyloid-like nanostructures (Kwan et al., 2006). However, our observed X-ray reection pattern has not been previously reported and appears to represent a new nanostructured form of wheat proteins. Rauscher et al. (2006) have discussed the role of proline and glycine in control of protein self-organization into elastomeric or
Fig. 4. Diffraction patterns obtained from dried SDS-insoluble glutenin proteins at 2 mg/ml, pH 7.4 for 105 days.
162
S.H. Mackintosh et al. / Journal of Cereal Science 49 (2009) 157162 Kipp, B., Belitz, H.-D., Seilmeier, W., Wieser, H., 1996. Comparative studies of high Mr subunits of rye and wheat. I. Isolation and biochemical characterisation and effects on gluten extensibility. Journal of Cereal Science 23, 227234. Kwan, A.H., Wineeld, R.D., Sunde, M., Matthews, J.M., Haverkamp, R.G., Templeton, M.D., Mackay, J.P., 2006. Structural basis for rodlet assembly in fungal hydrophobins. Proceedings of the National Academy of Sciences 103, 36213626. LeVine III, H., 1999. Quantication of beta-sheet amyloid bril structures with thioavin T. In: Wetzel, R. (Ed.), Methods in Enzymology 309: Amyloid, Prions, and Other Protein Aggregates. Academic Press, US, pp. 274284. Lew, E.J.-L., Kuzmicky, D.D., Kasarda, D.D., 1992. Characterization of low molecular weight glutenin subunits by reversed-phase high-performance liquid chromatography, sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and Nterminal amino acid sequencing. Cereal Chemistry 69, 505515. Lindsay, M.P., Skerritt, J.H., 1999. The glutenin macropolymer of wheat our doughs structurefunction perspectives. Trends in Food Science & Technology 10, 247253. Makin, O.S., Serpell, L.C., 2004. Structural characterisation of islet amyloid polypeptide brils. Journal of Molecular Biology 335, 12791288. Makin, O.S., Serpell, L.C., 2005. Structures for amyloid brils. FEBS Journal 272, 59505961. Makin, O.S., Atkins, E., Sikorski, P., Johansson, J., Serpell, L.C., 2005. Molecular basis for amyloid bril formation and stability. Proceedings of the National Academy of Sciences of the United States of America 102, 315320. Nilsson, M.R., Dobson, C.M., 2003. Chemical modication of insulin in amyloid brils. Protein Science 12, 26372641. Osborne, T.B., 1907. The proteins of the wheat kernel. In: Judd, Detweiller (Eds.), Publications of the Carnegie Institution Washington No. 84, pp. 1119. Washington, DC. Pawar, A.P., DuBay, K.F., Zurdo, J., Chiti, F., Vendruscolo, M., Dobson, C.M., 2005. Prediction of aggregation-prone and aggregation-susceptible regions in proteins associated with neurodegenerative diseases. Journal of Molecular Biology 350, 379392. Perutz, M.F., Finch, J.T., Berriman, J., Lesk, A., 2002. Amyloid bers are water-lled nanotubes. Proceedings of the National Academy of Sciences 99, 55915595. Rauscher, S., Baud, S., Miao, M., Keeley, F.W., Pomes, R., 2006. Proline and glycine control protein self-organization into elastomeric or amyloid-like brils. Structure 14, 16671676. Sanford, K., Kumar, M., 2005. New proteins in a materials world. Current Opinion in Biotechnology 16, 416421. Serpell, L.C., Fraser, P.E., Sunde, M., 1999. In: Wetzel, R. (Ed.), Methods in Enzymology 309: Amyloid, Prions, and Other Protein Aggregates. Academic Press, US, pp. 526536. Serpell, L.C., 2000. Alzheimers amyloid brils: structure and assembly. Biochimica et Biophysica Acta 1502, 1630. Shen, X., Mo, X., Moore, R., Frazier, S.J., Iwamoto, T., Tomich, J.M., Sun, X.S., 2006. Adhesion and structural properties of protein nanomaterials containing hydrophobic and charged amino acids. Journal of Nanoscience and Nanotechnology 6, 837844. Shewry, P.R., Halford, N.G., Belton, P.S., Tatham, A.S., 2002. The structure and properties of gluten: an elastic protein from wheat grain. Philosophical Transactions B 357, 133142. Stathopulos, P.B., Scholz, G.A., Hwang, Y.M., Rumfeldt, J.A., Lepock, J.R., Meiering, E.M., 2004. Sonication of proteins causes formation of aggregates that resemble amyloid. Protein Science 13, 30173027. Sugiyama, T., Rafalski, A., Peterson, D., Soll, D., 1985. A wheat HMW glutenin subunit gene reveals a highly repeated structure. Nucleic Acids Research 13, 87298737. Tatham, A.S., 1995. The structures of wheat proteins. In: Schoeld, J.D. (Ed.), Wheat Structure, Biochemistry and Functionality. The Royal Society of Chemistry, Reading, U.K., pp. 5362. Tatham, A.S., Shewry, P.R., 2002. Comparative structures and properties of elastic proteins. Philosophical Transactions B 357, 229234. Wang, S.S., Carpenter, F.H., 1967. Kinetics of the tryptic hydrolysis of the oxidized B chain of bovine insulin. Biochemistry 6, 215224. Waterhouse, S.H., Gerrard, J.A., 2004. Amyloid brils in bionanotechnology. Australian Journal of Chemistry 57, 519523.
amyloid-like brils. Given that HMW-GS contain repeating sequences of proline, glycine and glutamine (Tatham and Shewry, 2002), it may well be that this novel structure is dependent on the interaction of these three amino acids and could shed light on the relationship between sequence, structure and elastomeric qualities of these intriguing proteins. Acknowledgements The authors thank Harald Dobberstein and Manfred Ingerfeld for assistance with TEM, Grant Pearce and Amol Pawar for assistance with the Zygreggator algorithm, and Drs Cait MacPhee, Kate McGrath, Sarah Meehan, and Margie Sunde for useful discussions. Funding was provided by the New Zealand Foundation for Research, Science and Technology via the New Economy Research Fund contracts C02X0204 (Advanced Biological Materials) and C02X0404 (Amyloid Fibrils). References
Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J., 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Research 25, 33893402. Anjum, F.M., Khan, M.R., Din, A., Saaed, M., Pasha, I., Arshad, M.U., 2008. Wheat gluten: high molecular weight glutenin subunits structure, genetics, and relation to dough elasticity. Journal of Food Science 72, R56R63. Ban, T., Yamaguchi, K., Goto, Y., 2006. Direct observation of amyloid bril growth, propagation, and adaptation. Accounts of Chemical Research 39, 663670. Bradford, M.A., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilising the principle of proteindye binding. Analytical Biochemistry 72, 248254. Brenner, S., Horne, R.W., 1959. A negative staining method for high resolution electron microscopy of viruses. Biochimica et Biophysica Acta 34, 103110. Chen, S., Berthelier, V., Hamilton, J.B., ONuallain, B., Wetzel, R., 2002. Amyloid-like features of polyglutamine aggregates and their assembly kinetics. Biochemistry 41, 73917399. Chiti, F., Dobson, C.M., 2006. Protein misfolding, functional amyloid, and human disease. Annual Review of Biochemistry 75, 333366. DeMarco, M.L., Silveira, J., Caughey, B., Daggett, V., 2006. Structural properties of prion protein protobrils and brils: an experimental assessment of atomic models. Biochemistry 45, 1557315582. Dobson, C.M., 1999. Protein misfolding, evolution and disease. Trends in Biochemical Sciences 24, 329332. Dobson, C.M., 2001. The structural basis of protein folding and its links with human disease. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 356, 133145. Gerrard, J.A., Fayle, S.E., Brown, P.K., Sutton, K.H., Simmons, L., Rasiah, I., 2001. Effects of microbial transglutaminase on the wheat proteins of bread and croissant dough. Journal of Food Science 66, 782786. Guilbert, S., Gontard, N., Morel, M.H., Chalier, P., Micard, V., Redl, A., 2001. Formation and properties of wheat gluten lms and coatings. In: Gennadios, A. (Ed.), Protein-based Films and Coatings. CRC Press, Boca Raton, pp. 69122. Gras, S., 2007. Amyloid brils: from disease to design. New biomaterial applications for self-assembling cross-b brils. Australian Journal of Chemistry 60, 333342. Kasarda, D.D., Bernardin, J.E., Thomas, R.S., 1967. Reversible aggregation of a-gliadin to brils. Science 155, 203205. Kim, J.R., Muresan, A., Lee, K.Y., Murphy, R.M., 2004. Urea modulation of b-amyloid bril growth: experimental studies and kinetic models. Protein Science 13, 28882898.

Journal of Cereal Science - v49 - Iss1

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Journal of Cereal Science - v49 - Iss1

Uploaded by

Copyright:

Available Formats

Volume 49, Issue 1, Pages 1-162 (January 2009)

IFC - Editorial Board Page IFC