ACADEMIE DE MONTPELLIER
Centre international d’études supérieures en sciences agronomiques
THESE Pour obtenir le grade de
Docteur du Centre International d’Etudes Supérieures en Sciences Agronomiques de Montpellier Formation doctorale : Biochimie, Chimie et Technologie des aliments Ecole doctorale : Sciences des Procédés - Sciences des Aliments
Pâtes alimentaires enrichies en légumineuse. Structuration des constituants au cours du procédé : Impact sur la qualité culinaire et les propriétés nutritionnelles des pâtes par
Maud PETITOT Le 10 Septembre 2009
JURY Mme Ambrogina Pagani, UNIMI, DiSTAM, Milan
Rapporteur
Mme Santé-Lhoutellier Véronique, INRA, Saint-Genès Champanelle
Rapporteur
Mme Minier Chantal, Directrice R&D du C.R.E.C.E.R.P.A.L, Marseille
Examinateur
M. Avignon Antoine, Professeur, CHU Lapeyronie, Montpellier
Examinateur
Mme Micard Valérie, Professeur, Montpellier SupAgro
Directeur de thèse
Remerciements J’adresse mes sincères remerciements à … …Stéphane Guilbert, Professeur à Montpellier SupAgro, pour m’avoir accueillie au sein de l’UMR « Ingénierie des Agropolymères et Technologies Emergentes ». …Valérie Micard, Professeur à Montpellier SupAgro, pour m’avoir fait l’honneur de me confier cette thèse, pour sa confiance et sa disponibilité. …Toutes les personnes du centre de recherche européen sur les céréales (Crecerpal, Panzani) et en particulier Chantal Minier, Directrice R&D, Isabelle Arekion, Responsable R&D produits céréaliers et Laurence Boyer, technicienne, pour leur collaboration et leur accueil chaleureux à chacune de mes visites. …Les membres du jury, Véronique Santé-Lhoutellier, Chargée de recherche à l’INRA de Saint Genès-Champanelle et Ambrogina Pagani, Professeur à l’université de Milan, pour m’avoir fait l’honneur d’être rapporteurs de ma thèse ; mais également Antoine Avignon, Professeur au CHU Lapeyronie de Montpellier et Chantal Minier, Directrice R&D de Panzani à Marseille, pour avoir accepté d’examiner mon travail de thèse. …Brigitte Bouchet, Fabienne Guillon et Véronique Planchot de l’INRA de Nantes, pour avoir accepté de suivre le déroulement de mon travail de thèse. …Tous les partenaires du Projet PASTALEG : Isabelle Crenon, Dominique Jourdheuil Rahmani, Martine Champ, et Colette Larré. Je remercie tout particulièrement Chantal Brossard, pour sa précieuse collaboration et son amabilité, sans oublier Guilherme Moretto, Bruno Chauvet et Julien Maillard qui ont participé activement à la production des pâtes et à leur digestion ! …Béatrice Chabi, Barbara Vernus et Gaëlle Viennois, pour leur aide précieuse à la préparation et l’observation de mes pâtes sous le microscope ! …Tous mes collègues de l’UMR IATE et en particulier ceux qui m’ont permis d’avancer dans mes recherches : Marie-Hélène Morel, Cécile Barron, Bernard Cuq, et Joël Abecassis, mais également Joelle Bonicelle, Thérèse Marie Lasserre et Yannick Mellerin, pour m’avoir assistée dans mes expérimentations et pour m’avoir fait part de leur savoir faire, sans oublier Christophe Duprat et Anne-Marie Delmont, deux secrétaires hors pairs ! …Mes très très très chères amies: Mimi, Mariana, Gabi, Youna, Gisela et Rallou pour leur joie de vivre, leur humour et leur gentillesse. …Ma famille, pour son amour et son soutien, avec une dédicace spéciale au Roi des recettes au potimarron !
Sommaire
Sommaire
Sommaire / Table of contents
Liste des abréviations et symboles............................................................................................... 1 Liste des tableaux.......................................................................................................................... 2 Liste des figures............................................................................................................................. 5 Introduction générale ................................................................................................................. 11
Chapitre 1. Synthèse bibliographique....................................................................................... 19 Partie 1. Publication 1: Structuring of pasta components during processing: Impact on starch and protein digestibility and allergenicity. A review........................................................................... 19 1. Introduction................................................................................................................... 20 2. Structuring of pasta components during processing ..................................................... 21 2.1. Mixing.................................................................................................................... 22 2.2. Forming.................................................................................................................. 23 2.3. Drying .................................................................................................................... 26 2.4. Cooking.................................................................................................................. 28 3. Microstructure of cooked pasta: impact on starch and protein digestibility and potential allergenicity....................................................................................................................... 32 4. Conclusion .................................................................................................................... 37 Partie 2. Publication 2: Fortification of pasta with legume flour: processing ease, cooking quality, structure and nutritional properties ............................................................................................... 39 1. Introduction................................................................................................................... 40 2. Composition of durum wheat and some grain legumes................................................ 40 2.1. Global composition................................................................................................ 40 2.2. Proteins .................................................................................................................. 41 2.3. Starch ..................................................................................................................... 42 3. Nutritional value of wheat and grain legumes and their food products ........................ 45 3.1. Nutritional value of wheat and wheat products ..................................................... 45 3.2. Nutritional value of legumes.................................................................................. 48 4. Fortification of pasta with legume flour ....................................................................... 51 4.1. Technological issues .............................................................................................. 51 -i-
Sommaire 4.2. Cooking and organoleptic properties of legume fortified pasta .............................52 4.3. Nutritional properties of legume fortified pasta .....................................................53 5. Conclusion.....................................................................................................................55
Chapitre 2. Matériels et méthodes .............................................................................................59 1. Caractérisation des matières premières .........................................................................59 1.1. Teneur en eau des matières premières (publications 3, 4, 5 et 6)...........................59 1.2. Composition chimique des matières premières (publications 4, 5 et 6).................59 1.3. Granulométrie des matières premières et de l’amidon (publications 4 et 5)..........60 1.4. Solubilité des protéines et coagulation à la chaleur (publication 5).......................60 1.5. Propriétés d’hydratation des matières premières (publication 4) ...........................61 2. Fabrication des pâtes alimentaires.................................................................................61 2.1. Fabrication des pâtes alimentaires au blé dur (publications 3, 4, 5).......................61 2.2. Fabrication des pâtes enrichies en légumineuse (publications 4, 5, 6)...................64 2.3. Propriétés d’agglomération des matières premières au cours du malaxage (publication 4) ...............................................................................................................64 3. Tests de cuisson.............................................................................................................64 3.1. Temps optimal de cuisson (publications 3, 4 ,5 ,6) ................................................64 3.2. Cinétique d’absorption d’eau au cours de la cuisson (publications 5 et 6) ............65 3.3. Mesure de l’eau absorbée dans les pâtes cuites (OCT + 1min) (publication 4) .....66 3.4. Pertes à la cuisson (publication 4) ..........................................................................66 4. Caractérisation des principaux constituants des pâtes...................................................66 4.1. Caractérisation de l’amidon....................................................................................66 4.2. Caractérisation de l’état d’agrégation des protéines (publications 3, 5 et 6)..........68 5. Caractérisation des pâtes ...............................................................................................70 5.1. Porosité des pâtes sèches (publications 5 et 6).......................................................70 5.2. Couleur des pâtes sèches et des pâtes cuites (publication 4)..................................70 5.3. Rhéologie des pâtes cuites (publications 3, 4, 5, 6)................................................71 5.4. Microsctructure des pâtes cuites (publications 3, 5 et 6)........................................72 5.5. Analyse sensorielle des pâtes (publication 4).........................................................76 5.6. Propriétés nutritionnelles (publications 3, 5 et 6)...................................................76
- ii -
Sommaire Chapitre 3. Pâtes alimentaires au blé dur. Influence des barèmes de séchage sur leur structure et leurs propriétés nutritionnelles............................................................................. 81 Chapitre 3 : INTRODUCTION..................................................................................................... 81 Publication 3: Modification of pasta structure induced by high drying temperatures. Effects on the in vitro digestibility of protein and starch fractions and the potential allergenicity of protein hydrolysates .................................................................................................................................. 83 1. Introduction................................................................................................................... 84 2. Material and Methods ................................................................................................... 85 2.1. Pasta manufacturing............................................................................................... 85 2.2. DSC measurements on dried and cooked spaghetti............................................... 85 2.3. Protein extractability of dried and cooked spaghetti.............................................. 86 2.4. Microscopic characterisation of the protein network in cooked spaghetti ............ 87 2.5. Rheological properties of cooked spaghetti........................................................... 89 2.6. In vitro digestion of cooked and minced spaghetti ................................................ 90 2.7. IgE dot blotting and competitive ELISA ............................................................... 91 3. Results........................................................................................................................... 93 3.1. Effect of the drying profile on the constituents of dried spaghetti ........................ 93 3.2. Effect of the drying profile on cooked spaghetti ................................................... 96 4. Discussion ................................................................................................................... 109 Chapitre 3: CONCLUSIONS ...................................................................................................... 112
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire ..................................................................................................................................................... 115 Chapitre 4 : INTRODUCTION................................................................................................... 115 Publication 4 : Fortification of pasta with split pea and faba bean flours : pasta processing and quality evaluation........................................................................................................................ 117 1. Introduction................................................................................................................. 118 2. Materials and Methods................................................................................................ 119 2.1. Raw materials....................................................................................................... 119 2.2. Chemical composition of semolina and flours..................................................... 119 2.3. Particle size distribution of semolina and flours.................................................. 119 2.4. Ability of legume flours to be processed into pasta............................................. 119 2.5. Spaghetti cooking quality .................................................................................... 121 - iii -
Sommaire 2.6. Colour of dry and cooked spaghetti......................................................................122 2.7. Rheological properties of cooked spaghetti .........................................................122 2.8. Sensory evaluation................................................................................................123 2.9. Statistical analyses................................................................................................123 3. Results and discussion.................................................................................................124 3.1. Chemical composition of semolina and flours .....................................................124 3.2. Ability of legume flours to be processed into pasta .............................................124 3.3. Cooking quality evaluation...................................................................................128 3.4. Colour of dry and cooked spaghetti......................................................................129 3.5. Rheological properties..........................................................................................131 3.6. Sensory evaluation................................................................................................132 4. Conclusion...................................................................................................................134 Chapitre 4. CONCLUSIONS ......................................................................................................135
Chapitre 5. Pâtes enrichies en légumineuse. Impact de la formulation et des traitements technologiques sur la structure des pâtes et la digestibilité de l’amidon..............................139 Chapitre 5: INTRODUCTION ....................................................................................................139 Partie 1. Publication 5: Structure of pasta fortified pasta with split pea or faba bean flour: impact on starch digestibility ..................................................................................................................141 1. Introduction .................................................................................................................142 2. Materials and Methods ................................................................................................143 2.1. Chemical composition of durum wheat semolina and legume flours. Protein solubility and heat coagulation of soluble proteins .....................................................143 2.2. Pasta manufacturing .............................................................................................144 2.3. Porosity of dry spaghetti.......................................................................................144 2.4. Water absorption kinetic during cooking .............................................................144 2.5. Microstructure ......................................................................................................145 2.6. Protein extractability during pasta processing......................................................147 2.7. DSC analysis of starch in dry and cooked spaghetti ............................................148 2.8. In vitro starch digestibility of cooked spaghetti ...................................................149 3. Results .........................................................................................................................149
- iv -
Sommaire 3.1. Chemical composition, solubility in water and heat coagulation of soluble proteins in durum wheat semolina and legume flours .............................................................. 149 3.2. Porosity of dry spaghetti ...................................................................................... 151 3.3. Water absorption kinetic during cooking............................................................. 151 3.4. Microstructure...................................................................................................... 153 3.5. Protein reticulation during processing of durum wheat, split pea and faba bean pasta ............................................................................................................................ 160 3.6. DSC analysis in dry and cooked spaghetti........................................................... 164 3.7. Starch digestibility ............................................................................................... 165 4. Discussion and conclusion.......................................................................................... 166 Partie 2. Publication 6: Legume fortified pasta. Impact of drying and precooking treatments on pasta structure and inherent starch digestibility.......................................................................... 171 1. Introduction................................................................................................................. 172 2. Materials and methods ................................................................................................ 173 2.1. Chemical composition of durum wheat semolina and legume flours.................. 173 2.2. Pasta manufacturing............................................................................................. 173 2.3 Porosity of dry pasta ............................................................................................. 174 2.4. Determination of optimal cooking time and hydration kinetic during cooking... 174 2.5. Microstructure of cooked spaghetti ..................................................................... 175 2.6. Protein size distribution in dry and cooked spaghetti .......................................... 176 2.7. DSC analyses of dried and cooked spaghetti....................................................... 177 2.8. In vitro starch digestibility of cooked pasta......................................................... 178 3. Results......................................................................................................................... 178 3.1. Chemical composition ......................................................................................... 178 3.2. Porosity of dry spaghetti ...................................................................................... 179 3.3. Optimal cooking time and spaghetti hydration kinetic during cooking............... 179 3.4. Microstructure of cooked split pea pasta ............................................................. 181 3.5. Protein size distribution in dry and cooked spaghetti .......................................... 184 3.6. DSC measurements in dry and cooked spaghetti................................................. 186 3.7. In vitro starch digestibility of cooked spaghetti................................................... 187 4. Discussion and conclusion.......................................................................................... 188 Chapitre 5. CONCLUSIONS ................................................................................................. 191
-v-
Sommaire Chapitre 6. Discussion générale ...............................................................................................195 1. Objectifs et démarche employée .................................................................................195 2. Synthèse et discussion des principaux résultats obtenus.............................................196
Conclusions et perspectives ......................................................................................................211 Références bibliographiques.....................................................................................................217 Travaux relatifs à cette étude ...................................................................................................231 Publications dans des revues internationales à comité de lecture ...................................231 Communications orales à des congrès scientifiques .......................................................231 Communications affichées à des congrès scientifiques...................................................232
- vi -
Abréviations
Liste des abréviations et symboles
∆H:
geltinization enthalpy
∆Tr
temperature range
Tc
conclusion temperature
To:
onset temperature
Tp
peak temperature
CSLM:
confocal laser scanning microscopy
db:
dry basis
DH:
degree of hydrolysis
DSC:
differential scanning calorimetry
DTE:
dithioerythritol
DW:
durum wheat
ESEM:
environmental scanning electron microscopy
FB:
faba bean
GI:
glycemic index
HT:
high temperature
Ig-E:
immunoglobuline E
LT:
low temperature
LYO:
lyophilised pasta
OCT:
optimal cooking time
PCA:
principal component analysis
PreC:
precooked pasta
RAG:
rapidly available glucose
RS:
resistant starch
SE-HPLC:
size-exclusion high performance liquid chromatography
SDS:
sodium dodecyl sulfate
SP:
split pea
TPA:
texture profile analysis
VHT:
very high temperature
VHT.LM:
very high temperature applied at low moisture content
-1-
Tableaux
Liste des tableaux
Chapitre 1 Table 1.1.
Protein solubility and DSC analysis of starch during pasta processing (p 25).
Table 1.2.
Global composition of durum wheat and some grain legumes (% db) (p 41).
Table 1.3.
Solubility of wheat and some legume proteins (p 42).
Table 1.4.
Thermal characteristic (DSC) of durum wheat and some legume starches: onset (To), peak (Tp) and conclusion (Tc) temperatures (°C) (p 45).
Table 1.5.
Glycemic index (GI) of wheat products, other cereal products or legumes, with glucose or bread as the standard food (p 46).
Table 1.6.
α-galactosides content of some grain legumes (% db) (p 49).
Table 1.7.
Amino-acid scores of durum wheat semolina and grain legumes, based on FAO/WHO/UNU (1985) pattern for a pre-schooled child (2-5 years) (p 50).
Chapitre 3 Table 3.1.
DSC measurements obtained from the first endothermic peak of DSC on dried spaghetti dried with one of the four drying profiles (LT, HT, VHT.LM, or VHT). Means (n=4) with the same superscript within a column are not significantly different (P>0.05). To, onset temperature; Tp, peak temperature; Tc; conclusion temperature, ∆Tr = Tc-To, temperature range, ∆H gelatinization enthalpy (p 93).
Table 3.2
Textural characteristics of cooked pasta from Texture Profile Analysis (TPA) and tensile test. Means (n=10) with the same superscript within a column are not significantly different (P>0.05) (p 102).
Table 3.3.
Statistical analysis of the variations of the degree of hydrolysis of proteins (DH) and of the percentage of alcohol-soluble dextrins (PASD) during the bucco-gastric and duodenal phases. Results of the ANOVA and subsequent LSD test (p 104).
Table 3.4.
Results of competitive ELISA with digestion juices from pasta and a pool of sera from allergic patients to wheat. Percentage of inhibition obtained with digestion juices at the end of the gastric phase (5 minutes of hydrolysis by α-amylase and 3 hours by pepsin) and at the end of the intestinal phase (5 minutes of hydrolysis by α-amylase, 3 hours by pepsin and 3 hours by pancreatin) are presented. Means (n=2) with the same superscript by protein are not significantly different (P > 0. 05) (p 108).
-2-
Tableaux
Chapitre 4 Table 4.1.
Chemical composition of durum wheat semolina, split pea flour, faba bean flour and blends of 35% legume flour and 65% durum wheat semolina. Protein, starch, fiber, lipid and ash contents are expressed in g/100g (db); vitamin and mineral content is expressed in mg/100g (db) (p 125).
Table 4.2.
Granulation properties of durum wheat (DW) semolina, and blends made from 65% durum wheat semolina and 35% split pea (SP) flour or faba bean (FB) flour and pasta processing control (p 126).
Table 4.3.
Cooking quality of pasta made from durum wheat (DW) semolina (control) or a blend made from 65% DW semolina and 35% spli pea (SP) flour or faba bean (FB) flour, and dried at low (LT), high (HT), or very high temperature applied at the end of the drying cycle (VHT.LM) (p 130).
Table 4.4.
Sensory evaluation of cooked (OCT+1) pasta made from durum wheat (DW) semolina (control) or a blend of 65% DW semolina and 35% of legume flour split pea (SP) flour or faba bean (FB) flour, and dried at low (LT), high (HT), or very high temperature applied at the end of the drying cycle (VHT.LM) (p 133).
Chapitre 5 Table 5.1.
Chemical composition, protein solubility (in water, pH 6,3) and heat coagulation of soluble proteins at 100°C of durum wheat (DW) semolina, split pea (SP) flour, faba bean (FB) flour and blends of 35% legume flour and 65% durum wheat semolina (p 150).
Table 5.2.
Water absorption and in vitro starch digestibility of durum wheat (DW) pasta, split pea (SP) pasta and faba bean (FB) pasta cooked at their optimal cooking time (OCT) + 1min (p 152).
Table 5.3.
Two way analysis of variance and subsequent LSD test of the scores on principal component 1 (PC1) and principal component 2 (PC2) of the principal component analysis of the erosion-dilation curves for durum wheat (DW) pasta, split pea (SP) pasta and faba bean (FB) pasta (p 158).
Table 5.4.
DSC measurements obtained from the first and the second endothermic peak of DSC on durum wheat (DW) semolina, blends of 35% split pea (SP) or faba bean (FB) flour and 65% DW semolina, and dried spaghetti. Means (n=3) with the same superscript within a column are not significantly different (P > 0.05). To: onset temperature; Tp: peak temperature; Tc: conclusion temperature, ∆Tr = Tc-To: temperature range, ∆H: enthalpy (p 164).
Table 5.5.
Chemical composition of durum wheat (DW) semolina, split pea (SP) flour, faba bean (FB) flour and blends of 35% legume flour and 65% durum wheat semolina (p 178).
-3-
Tableaux
Table 5.6.
Two way analysis of variance and subsequent LSD test on optimal cooking time, water absorption and in vitro rapidly available glucose (RAG) of split pea (SP) pasta and faba bean (FB) pasta cooked at their optimal cooking time (OCT) + 1min (p 180).
Table 5.7.
Two way analysis of variance and subsequent LSD test on scores obtained from principal component (PC) analysis conducted on erosion-dilation curves of split pea pasta images (p 182).
Table 5.8.
Two-way analysis of variance on DSC measurement obtained from the first endothermic transition of dried pasta. To: onset temperature; Tp, peak temperature; Tc; conclusion temperature, ∆Tr = Tc-To, temperature range, ∆H gelatinisation enthalpy (p 187).
Chapitre 6 : Discussion générale Table 6.1.
Glucose rapidement disponible (RAG) issu des analyses de digestion in vitro calculés en % des glucides disponibles (p 201).
Table 6.2.
Pourcentage des protéines solubles dans le SDS, solubles dans le DTE après sonication, et insolubles dans les pâtes cuites au blé dur (DW), au pois cassé (SP) ou à la fève (FB), ayant été séchées à basse température 55°C (LT) ou très haute température 90°C en fin de cycle (VHT.LM) (p 205).
-4-
Figures
Liste des figures
Introduction générale Figure 1.
Démarche générale du travail de thèse (page 13).
Chapitre 1 Figure 1.1.
Scanning electron microscopy of A) Semolina (Petitot, not published) B) freshly extruded pasta (Tudorica et al., 2002), C) and D) Surface and cross section of dried spaghetti (85°C) (Sadeghi & Bhagya, 2008) (page 23).
Figure 1.2.
Schematic representation of the temperature-water content state diagram for the main components in wheat flour (starch and gluten proteins). Tg is the glass transition temperature, Tm is the melting point, Tgelat is the starch gelatinisation temperature, and Tr is the minimum temperature for protein thermosetting (Cuq et al, 2003) (page 29).
Figure 1.3.
Low voltage scanning electron microscopy (A, B, C, D) and bright field light microscopy (E, F,G) of cooked HT-dried spaghetti. A: Surface of spaghetti, B-G: internal structure of spaghetti. From the left to the right: from the centre to the periphery of spaghetti. E-G: fast green-iodine staining: proteins stain green while starch granules stain blue-brown. Adapted from Heneen & Brismar (2003) (page 31).
Figure 1.4.
Confocal laser scanning microscopy of HT-dried cooked pasta in the central (A), intermediate (B) and external (C) regions of pasta strand. Proteins were stained with fuchsin acid. White areas correspond to proteins and dark areas can be assimilated to starch granules (Petitot, not published) (page 31).
Figure 1.5.
Chemical structure of amylose and amylopectin (page 43).
Figure 1.6.
Wheat and legume starch granules observed by scanning electron microscopy (Petitot, not published) (page 43).
Figure 1.7.
Schematic representation of hydrothermal transformations of a starch granule (page 44).
Figure 1.8.
Schematic representation of glycemic index calculation (page 46).
-5-
Figures
Chapitre 2 Figure 2.1.
Présentation de la presse pilote et de ses accessoires (page 62).
Figure 2.2.
Diagrammes de séchage appliqués au cours du séchage des pâtes dans le séchoir pilote. HR : humidité relative ; T : température (page 63).
Figure 2.3.
Illustration du test de la ligne blanche utilisé pour la détermination du temps optimal de cuisson. A : pâte non cuite, B : pâte cuite, C : pâte cuite au temps optimal de cuisson (page 65).
Figure 2.4.
Exemple de thermogramme obtenu par analyse enthalpique différentielle (DSC) (page 67).
Figure 2.5.
Représentation du test en extension (page 71).
Figure 2.6.
Représentation du test TPA (page 72).
Figure 2.7.
Exemple de courbe d’érosion-dilatation obtenu après analyse d’image des pâtes au blé dur cuites (page 75).
Chapitre 3 Figure 3.1.
Peak areas of SE-HPLC elution profiles of SDS-soluble, DTE-soluble and unextractable protein fractions in semolina and dried pasta (A) and in cooked pasta (B). Means (n=3) with the same superscript within a graph (A or B) are not significantly different (P>0.05) (page 95).
Figure 3.2.
SE-HPLC elution profiles (means=3) of SDS-soluble (A) and DTE-soluble (B) protein fractions in cooked pasta dried with one of the four drying profiles (page 97).
Figure 3.3.
Typical images (acquired by CLSM) of the LT cooked pasta in the central (A), intermediate (B), and external (C) zones and granulometric curves (D) corresponding to the three images (page 99).
Figure 3.4.
Principal component analysis of the granulometric curves (A) and first loading corresponding to principal component 1 (B) (page 100).
Figure 3.5.
SE-HPLC profiles at 220 nm of digestion juices of LT-pasta in the course of in vitro digestion (A) and (B) loadings and (C) first plan of the Principal Components Analysis (PCA) of SE-HPLC profiles of pasta at the end of buccal phase, gastric phase, after bucco-gastric phase + 30 min of pancreatic phase and after bucco-gastric-phase +3h of pancreatic phase (page 106).
-6-
Figures
Chapitre 4 Figure 4.1. Particle size distribution of durum wheat semolina and legume flours (page 127). Figure 4.2. Rheological properties of cooked durum wheat pasta and pasta fortified with 35% of split pea or faba bean flour, dried at low (LT), high (HT), or very high temperature applied at the end of the drying cycle (VHT.LM). Different letters above the bars within each graph indicate significant difference (P60°C) (Weegels & Hamer, 1998). However, as the dough is transferred in the extrusion worm, pressure builds up and the dough temperature rises locally. Structural transformations are therefore a consequence of both mechanical (shearing stress) and thermal forces involved during extrusion (Kruger et al., 1996). Concerning the starch fraction, mechanical forces can lead to moderate damage to starch granules (Table 1.1) (Icard-Vernière, 1999; Lintas & d' Appolonia, 1973). A local increase in temperature (>60°C) due to mechanical forces can also lead to starch gelatinisation. DSC measurements revealed a lower gelatinisation enthalpy of extruded pasta compared to semolina (Vansteelandt & Delcour, 1999; Zweifel, Conde-Petit & Escher, 2000) (Table 1.1). A reduced gelatinisation enthalpy could be explained by the presence of some gelatinised starch granules and/or some damaged starch granules which require less energy to melt (Biliaderis, 1990).
- 24 -
Chapitre 1 : Synthèse bibliographique
Table 1.1. Protein solubility and DSC analysis of starch during pasta processing End of Extruded Semolina mixing Pasta Solubility in SDS (% of total protein content)
Solubility in acetic acid
92.9 f
96.3 f
71.3 a 75.2 b
(% of total protein content)
88.9 f
66.2
b
LT < 55°C
70.0 g
63.0 g
66.9 a 66.0 b
64.8 a
22-28 g 63.4 a
15.0 g
46.7 a
65.0 c
10.0 c 50.0
e
Gelatinization enthalpy (J/g db)
Dried pasta HT HT VHT VHT Cooked pasta* 60-70°C 80-85°C 90-110°C 180°C
4.8 6.2 g 12.4 h 10.2 i 5.8 j
e
4.9 g 12.4 5.4
h
j
2.7 4.9 g
3.3
d
35.0
d
e
4.4-4.9 g h
9.6 i 4.9 j
12.2 9.9 i
9.9 i 5.2-6.8 j
Adapted from a Aktan & khan (1992), b Dexter & Matsuo (1977), c Dexter & Matsuo (1979), d de Zorzi et al. (2007), e Güler et al. (2002), f Icard-Verniere (1999), g Petitot et al. (2009c), h Vansteelandt & Delcour (1998), iYue et al. (1999), j Zweifel et al. (2000) When several varieties of durum wheat were tested, average data were calculated. * Data are given for cooked pasta, whatever the drying profile. LT: Low Temperature, HT: High Temperature, VHT: Very high temperature OCT: optimal cooking time. Optimal cooking time is indicated when the white core of the pasta disappeared when squeezed between two glass plates (Approved method 66-50 AACC 2000).
I.2.2. Sheeting
CLSM observations of freshly sheetd durum wheat pasta revealed that the proteins are closely associated with starch granules. With increasing sheeting passes (from 3 to 45 passes), the proteins and starch granules became distributed more uniformly throughout the dough. At the molecular scale, it induced a higher glutenin solubility in SDS due to protein disaggregation and depolymerisation (Kim et al., 2008). A moderate damage of starch was also reported (Zardetto & dalla Rosa, 2005).
I.2.3. Extrusion versus Sheeting
The comparison between the sheeting and the extrusion process dough remains difficult because of the lack of comparative studies. Pagani, Resmini & Dalbon (1989) studied the microstructure of fresh roll-sheeted and extruded pasta prepared from the same blend of common
- 25 -
Chapitre 1 : Synthèse bibliographique
wheat flour (T. aestivum). In freshly extruded spaghetti, the protein matrix looked discontinuous with protein aggregates unevenly distributed among the starch granules, whereas in fresh rollsheeted pasta, a compact and continuous protein network was observed. More recently, Fardet, Baldwin, Bertrand, Bouchet, Gallant & Barry (1998a) characterised the protein network of cooked sheeted lasagne and cooked extruded spaghetti, both made from durum wheat semolina. Lasagne exhibited a more porous and open protein/starch network and completely gelatinised starch granules in the centre of the strand. In contrast, spaghetti was characterised by a more compact structure with some ungelatinised starch granules in the central region. The contrasted results of Pagani et al. (1989) and Fardet et al. (1998a) could be mainly explained by the different raw materials used in both studies (common wheat flour vs. durum wheat semolina). At a lower scale, starch was found to be damaged by sheeting to a lesser extent than during extrusion (Zardetto et al., 2005), due to the lower “intrinsic stress” (lower temperatures and pressures and a shorter processing time) (Carini, Vittadini, Curti & Antoniazzi, 2009).
2.3. Drying
Since the 1970s, technological improvements in the pasta making process have induced an increase in drying temperatures: from low temperature (LT) (40-60°C / 70-80% relative humidity (RH) / 18-28 h), to high temperature (HT) (60-84°C/ 74-82% RH / 8-11 h), and very high temperature (VHT) (>84°C / 74-90% RH / 2-5 h), thus leading to shorter drying times and improved hygienic standards (Pollini, 1996). Moreover, the use of higher temperatures was also beneficial for the overall cooking quality of the final product, with higher firmness, lower stickiness and lower cooking loss (Aktan & Khan, 1992; Zweifel, Handschin, Escher & CondePetit, 2003). These improved properties are the results of structural modifications within pasta: depending on the temperature-moisture conditions applied during drying, physicochemical modifications (protein denaturation and starch swelling) occur in a different extent. When viewed under SEM, the surface of dried pasta presents numerous starch granules of different sizes (Figure 1.1 C) (Cunin et al., 1995; Sadeghi & Bhagya, 2008). In addition, they appear to be associated with a protein film. Some cracks and small holes are apparent in the protein matrix at the surface (Figure 1.1 C). This is probably partly due to surface tension in spaghetti dough during drying and partly due to shrinkage during sample preparation for microscopic observations (Sadeghi & Bhagya, 2008). The internal structure of dry pasta is characterised by starch granules deeply embedded in a protein matrix (Figure 1.1 D) (Cunin et al., 1995; Sadeghi et al., 2008). - 26 -
Chapitre 1 : Synthèse bibliographique
At the molecular level, up to a drying temperature of 70°C, minor changes in the extent of protein denaturation in pasta is observed: about 65% of proteins are soluble in acetic acid (Aktan et al., 1992; Dexter et al., 1977), and 63-70% are soluble in SDS (Petitot et al., 2009c) (Table 1.1). At higher temperatures, the protein solubility decreases drastically: at 90°C, only about 50% of proteins are soluble in acetic acid (Aktan et al., 1992; De Zorzi, Curioni, Simonato, Giannattasio & Pasini, 2007) and about 25% of proteins in SDS (Petitot et al., 2009c), which represents a 20–60% decrease in protein solubility compared to LT drying. Lamacchia et al. (2007) characterised proteins of dried pasta by SE-HPLC and observed a progressive decrease in the small and large monomeric proteins and an increase in the molecular size of large polymeric proteins as the drying temperature increased from 60 to 90°C. This was accompanied by an increasing amount of total unextractable polymeric protein (UPP) (20–100%), i.e. proteins that were not soluble in SDS or after sonication. Singh, Donovan, Batey & MacRitchie (1990) postulated that sonication causes shear degradation of disulphide bonds that connect glutenin subunits together. Sonication would not affect other types of covalent linkages (peptide bonds). This means that protein aggregates stabilised by irreversible protein interactions probably occur through interpeptide cross-linking or Maillard-type reactions, as observed by De Zorzi et al. (2007) and Petitot et al. (2009c) in pasta dried at VHT or by Micard, Morel, Bonicel Guilbert (2001) in pure gluten treated under severe conditions (130°C, 20% moisture content). However, disulphide bonds remain the main links formed during pasta drying. According to Fardet, Samson, Aubled, Morel & Abecassis (1996), glutenins are highly heat sensitive: at 80°C, they form intermolecular disulphide bonds and become insoluble in thermally treated fresh pasta (29% moisture content). At higher temperatures (>80°C), gliadins are also involved and form disulphide bonds with the glutenin complex. This is in accordance with studies carried out on pure gluten. Weegels, Verhoek, Degroot & Hamer (1994) studied the reactivity of gluten and found that major changes occurred when gluten was heated at 80°C and for a moisture content of 25-30% (w/w), which can be compared to the conditions applied during pasta drying at high temperature. A decrease in protein solubility was observed, affecting mainly the glutenin fraction. Gliadins appear to react at temperatures higher than 100° (Singh & MacRitchie, 2004). Glutenins are able to form both intermolecular and intramolecular disulphide bonds whereas gliadins form only intra-chain disulphide bonds that can participate in disulphide-sulphydryl interchange (Singh et al., 2004). However these observations concern the effect of temperature on gluten proteins in model system, and not in a complex matrix such as pasta where the interaction between pasta constituents has also to be considered.
- 27 -
Chapitre 1 : Synthèse bibliographique
Changes in pasta structure during drying cannot solely be explained according to the extent of protein denaturation. Changes in the starch fraction can also occur and are commonly analysed by polarized light microscopy (birefringence study), differential scanning calorimetry (DSC) and X-ray diffractometry, which reflect different levels of ordered structures (Zweifel et al., 2000). When viewed under polarized light, starch granules exhibited different birefringence levels in dry pasta. Most starch granules of LT dried pasta retained their birefringence (Altan & Maskan, 2005), whereas approximately 20% of starch granules of HT and VHT dried pasta either partially or completely lost their birefringence (Guler, Koksel & Ng, 2002). Moisture content and temperature conditions may have been reached locally to induce the gelatinisation of some starch granules. DSC analysis of the thermal properties of starch led to the detection of two endothermic transitions: the first one could be attributed to the gelatinisation of starch and the second corresponds to the reversible dissociation of pre-existing amylose-lipid complexes. Drying may promote partial melting of starch (Altan et al., 2005; Guler et al., 2002; Yue, RayasDuarte & Elias, 1999) and the formation of amylose-lipid complexes (Yue et al., 1999). However, the impact of the drying cycle on gelatinisation enthalpy has yet to be clarified. Yue et al. (1999) and Petitot et al. (2009c) did not find any significant difference among LT, HT and VHT dried pasta. In contrast, a higher gelatinisation temperature and enthalpy was found by Gület et al. (2002) and Zweifel et al. (2000) in pasta when increasing the drying temperature (Table 1.1). Zweifel et al. (2000) hypothesised that VHT drying may favour the molecular rearrangement of starch polymers. When analysed by X-ray diffractometry, no clear effect of the drying temperature in crystallinity was detected (Baiano et al., 2006; Guler et al., 2002; Zweifel et al., 2000).
2.4. Cooking
In durum wheat pasta, starch gelatinisation and protein coagulation are the main changes in pasta structure that occur during cooking. In the interspaces between granules, protein coagulation and interaction lead to a continuous and strengthened network, which traps the starch while the latter, by swelling and gelatinisation, occludes theses interspaces. Structural transformations in starch and proteins are in the same range of temperature and moisture conditions; proteins appear to react at a slightly lower moisture level (Figure 1.2) (Cuq, Abecassis & Guilbert, 2003). Both transformations are competitive (both component compete for water) and antagonistic (the swelling of starch granules is opposed to the formation of the protein network) (Pagani et al., 1986). The faster the starch swells, the slower the rate of protein - 28 -
Chapitre 1 : Synthèse bibliographique
interaction and the weaker the protein network inside the spaghetti. Both transformations are controlled by water penetration inside the pasta strand during cooking. Until the optimal cooking time, the water uptake rate depends on the ability of water to diffuse through the matrix; and on the melting kinetics of the crystalline domains (del Nobile & Massera, 2000). Water, which acts as a plasticizer and increases polymer mobility, penetrates concentrically toward the centre of spaghetti with cooking time. The presence of a moisture gradient inside the strand can be revealed by magnetic resonance imaging (MRI) (del Nobile et al., 2000; Horigane et al., 2006). The lower moisture content in the core of pasta strand may lead to a stronger competition for water between protein and starch granules. The water may be not evenly distributed among components. A higher hydration of proteins locally may induce the formation of the protein network before the swelling of the starch granules. As a consequence of this moisture gradient, a continuous change in the structure from the outer surface towards the core characterizes the internal framework of cooked pasta (Heneen & Brismar, 2003).
Figure 1.2. Schematic representation of the temperature-water content state diagram for the main components in wheat flour (starch and gluten proteins). Tg is the glass transition temperature, Tm is the melting point, Tgelat is the starch gelatinisation temperature, and Tr is the minimum temperature for protein thermosetting (Cuq et al, 2003).
- 29 -
Chapitre 1 : Synthèse bibliographique
Microscopic observations revealed that, on the smooth surface of cooked pasta, protein and starch are no longer distinguishable from one another (Figure 1.3 A), forming a thin film of about 1 µm thickness with some small cracks and open areas (Fardet, Hoebler, Baldwin, Bouchet, Gallant & Barry, 1998b; Heneen et al., 2003). After cooking pasta to its optimal cooking time, the internal structure of pasta can be divided into three concentric regions after cooking—an external region, an intermediate region, and a central region (Figures 1.3 and 1.4) (Cunin et al., 1995; Heneen et al., 2003; Petitot et al., 2009c). In the external region, starch granules are largely deformed, swollen and it is still difficult to differentiate them from proteins when viewed under SEM or BFLM (Figures 1.3 D and 1.3 G) (Fardet et al., 1998b; Heneen et al., 2003). The protein network is more clearly viewed under CLSM (Figure 1.4). Whatever the drying profile, the external region of cooked pasta is clearly distinguished from the intermediate and central regions. It is characterised by a higher amount of thin protein films surrounding larger starch granules (Petitot et al., 2009c). The intermediate region includes partly swollen granules embedded in a coagulated but dense protein network (Figures 1.3 C and F) (Fardet et al., 1998b; Heneen et al., 2003). The centre of the strand presents starch granules with a limited degree of gelatinisation (Figures 1.3 B and E), due to limited water absorption (Cunin et al., 1995).
At a smaller scale, the moisture content and temperature conditions during cooking are favourable for protein denaturation and aggregation. Indeed, a decrease in protein solubility in dilute acetic acid (Dexter & Matsuo, 1979) or in SDS (Petitot et al., 2009c) was observed during cooking (Table 1.1). Depending on the previous drying profile applied, the loss of protein solubility due to cooking can be more or less accentuated. For example, when comparing the solubility of proteins in pasta dried at different temperatures, it appeared that pasta dried at VHT underwent less marked changes during cooking compared to LT and HT dried pasta because most of the proteins were already aggregated during the drying step (Petitot et al., 2009c). Concerning the starch fraction, complete loss of birefringence is obtained when pasta is cooked to the optimal cooking time (Grzybowski & Donnelly, 1977; Marshall & Wasik, 1974). No gelatinisation endotherm is observed by DSC, suggesting the complete gelatinisation of starch in cooked pasta (Fardet et al., 1999; Petitot et al., 2009c).
- 30 -
Chapitre 1 : Synthèse bibliographique
Figure 1.3. Low voltage scanning electron microscopy (A, B, C, D) and bright field light microscopy (E, F,G) of cooked HT-dried spaghetti. A: Surface of spaghetti, B-G: internal structure of spaghetti. From the left to the right: from the centre to the periphery of spaghetti. E-G: fast green-iodine staining: proteins stain green while starch granules stain blue-brown. Adapted from Heneen & Brismar (2003).
Figure 1.4. Confocal laser scanning microscopy of HT-dried cooked pasta in the central (A), intermediate (B) and external (C) regions of pasta strand. Proteins were stained with fuchsin acid. White areas correspond to proteins and dark areas can be assimilated to starch granules (Petitot, not published).
- 31 -
Chapitre 1 : Synthèse bibliographique
Cooking is an essential step in which major structural transformations occur. However, these transformations are also dependent on those occurring during the previous processing steps. Indeed, the effects of some process parameters such as forming and drying are still detected after cooking. For example, forming seems to mainly affect the pasta porosity. Extruded spaghetti exhibits a slightly less porous and less open protein/starch network than laminated lasagne made from durum wheat semolina (Fardet et al., 1998b). Drying appears to be essential for maintaining the pasta structure, especially during overcooking: the protein network of LT (55°C) dried and then cooked pasta, viewed on CSLM images, is partly disrupted and has lost its continuity. The application of VHT (100°C) at a low moisture level preserves the microstructure of pasta, even after prolonged cooking (Fardet et al., 1998b). Moreover, the extent of starch swelling is lower in VHT (100°C) pasta, with the effect being more marked for pasta when the VHT phase is applied at a low moisture level (Zweifel et al., 2003). The final structure of pasta is therefore the result of successive changes occurring throughout the pasta making process and mainly affecting the starch and protein fractions. Modification of process parameters could therefore modify the pasta structure and impact its nutritional properties. Indeed, as discussed in the following section, the nutritional properties of pasta are closely linked to its structure.
3. Microstructure of cooked pasta: impact on starch and protein digestibility and potential allergenicity Pasta is a carbohydrate-based food (~70% starch) that is considered to be a source of slowrelease carbohydrates, therefore possessing a low glycaemic index (GI) (Jenkins, Wolever, Jenkins, Lee, Wong & Josse, 1983). The glycaemic index (GI) has been introduced in order to estimate the blood glucose response of foods after ingestion by humans. It is measured by the postprandial glycemic area of a test meal, expressed as the percentage of the corresponding area of the reference food (glucose of white bread). As human subjects would have to be recruited to measure GI, which is also time consuming, in vitro testing is commonly used as a faster method to predict in vivo GI (Casiraghi, Brighenti & Testolin, 1992; Fardet et al., 1999). This is based on in vitro measurement of the susceptibility of starch to digestive enzymes. Indeed, the glycaemic response and consequently the insulin demand appear to be closely related to the enzymic susceptibility of starch (Jenkins, Taylor & Wolever, 1982). The enzymic susceptibility of starch depends on the special organisation and structural state of pasta components. Hence, a
- 32 -
Chapitre 1 : Synthèse bibliographique
modification in the pasta structure through a pasta processing change could have an impact on the rate of starch degradation, i.e. the pasta GI. Numerous structural factors at the macroscopic, microscopic and molecular scales have been suggested to explain the slow rate of starch degradation in pasta :
At the macroscopic level (> 100 µm): the pasta size (i.e. vermicelli vs. spaghetti) and shape (i.e. macaroni vs. spaghetti), which depend on the forming step, seem to be of great importance. In spaghetti (large pasta size) starch was less susceptible to α-amylase (slower degradation rate) than in vermicelli (smaller pasta size). Higher surface to weight ratio of vermicelli vs. spaghetti may explain the higher accessibility of α-amylase to starch (Granfeldt et al., 1991a). The effect of the pasta shape was investigated by Wolever et al. (1986). These authors found that macaroni produced a significantly higher in vivo glucose response than spaghetti (68 ± 8 vs. 45 ± 8) but did not find any correlation between the GI and surface area of pasta. To assess the importance of the food structure with respect to the starch degradation rate, Grandfeldt, Björck & Hagander (1991b) compared the carbohydrate digestion of spaghetti, spaghetti porridge and bread made from the same raw material. Spaghetti produced significantly lower GI (GI=61) than bread (GI=100), as a result of different structures obtained after different manufacturing and cooking processes. Bread is characterised by an open structure with large holes and highly swollen starch granules whereas pasta presents a denser structure (Brennan, Symons & Tudorica, 2005). More interestingly, disintegration of spaghetti into a coarse ‘porridge’ significantly increased the glucose response (GI=73) compared to intact spaghetti (GI=61). This in accordance with Colonna et al. (1990) who demonstrated that grinding spaghetti led to a higher enzyme susceptibility of starch granules compared to intact spaghetti. The destruction of pasta structure probably facilitated the diffusion of α-amylase.
At the microscopic level (0.3–100 µm): in order to demonstrate the protective role played by the protein network and fibres, several authors have studied the in vitro enzymic susceptibility of starch after removal of the protein network or inclusion of fibres. The encapsulation of starch granules by fibres (Tudorica et al., 2002), and proteins (Colonna et al., 1990; Fardet et al., 1998b) was found to limit the accessibility of α-amylase to starch. Inclusion of high quantities of insoluble fibres (e.g. pea fibres) may disrupt the protein matrix, giving rise to a highly porous structure. The starch granules become more accessible, hence more susceptible to enzyme degradation (Tudorica et al., 2002). In contrast, inclusion of soluble fibers (e.g. guar) may induce the entrapment of starch granules within a viscous protein-fibre-starch network, which acts as a protective coat, leading to reduced glucose release (Tudorica et al., 2002). - 33 -
Chapitre 1 : Synthèse bibliographique
In order to highlight the protective role played by proteins, Colonna et al. (1990) removed the protein network of cooked pasta by Pronase® and found that starch degradation was increased. They also suggested that the application of high drying temperatures could lead to a high level of protein crosslinking, in turn leading to a more intense encapsulation of starch, thus decreasing its susceptibility to enzymes. Moreover, starch encapsulation by proteins could limit water absorption by starch granules, therefore limiting enzyme diffusion and decreasing α-amylolysis kinetics (Colonna et al., 1990). However, according to Fardet et al. (1998b) starch granules are not fully encapsulated by the protein network, whose porosity is high enough (0.5-40 µm) to allow α-amylase to diffuse freely. These authors suggested other hypotheses to explain the low GI of pasta: the tortuosity of the protein network which lengthens the pathway of α-amylases; and the presence of some high molecular weight starch polymers naturally resistant to enzymic attack.
At the molecular level (0.8–50 nm): the physical structure of starch, such as its degree of gelatinisation (Holm, Bjoerck & Eliasson, 1988a), retrogradation (Akerberg, Liljeberg & Björck, 1998), and amylose-to-amylopectin ratio (Akerberg et al., 1998; Holm et al., 1988a) also influence the rate of starch degradation. The more gelatinised the starch is, the more available it is to amylases (Holm et al., 1988a). However, microscopic observations (Pagani et al., 1986), DSC measurements (Petitot et al., 2009c) and X-ray diffractometry analyses (Colonna et al., 1990) have shown that starch granules are completely gelatinised in pasta cooked to its optimal cooking time. The consumption of a high amylose pasta meal (amylose content ≥ 39.6% of carbohydrate) (obtained by the replacement of 20% durum wheat flour by high amylose corn starch) induced a lower glucose and insulin response in humans compared to those who had consumed a moderate amylose pasta meal (amylose content 25.9% of carbohydrates) (obtained by the replacement of 20% durum wheat flour by normal amylose corn starch) (Hospers, van Amelsvoort & Weststrate, 1994). These authors suggested that a larger amount of slowly digestible amylose aggregates and resistant retrograded amylose starch may have been formed. Indeed, upon cooling, gelatinised starch retrogrades (especially amylose) and becomes resistant to hydrolysis (Svihus, Uhlen & Harstad, 2005). The gelatinisation of starch is not the factor limiting its degradation in cooked pasta. The use of high amylose starch and/or the presence a retrograded starch could induce lower starch hydrolysis in cooked pasta. Moreover, amylose is also prone to react with other food components such as lipids. The formation of theses complexes was shown to reduce the rate of amylolysis in rats (Holm et al., 1983), probably due to the reduction of surface contact between the enzyme
- 34 -
Chapitre 1 : Synthèse bibliographique
and its substrate and the decrease in starch swelling due to increasing hydrophobicity (Svihus et al., 2005). As shown in this review, the slow rate of starch degradation could be explained by the specific structure of pasta at the macroscopic, microscopic and molecular scales. This specific structure is the result of successive structural changes occurring throughout the pasta making process. However, it is hard to establish a definite link between pasta processing and carbohydrate digestion because of the lack of studies. It has been demonstrated, for example, that an increase in screw speed (150 to 250 rpm) or temperature (40°C to 70°C) during extrusion does not seem to affect in vitro starch hydrolysis in extruded cooked pasta (Fardet et al., 1999). In sheeted pasta, Kim et al (2008) suggested that several sheeting passes (3 to 45) contributed to the disruption of protein and starch interactions, leading to an increase in starch accessibility to
α-amylase. The impact of pasta drying has also been studied. VHT dryings were shown to decrease the in vitro digestibility of starch in cooked pasta (Casiraghi et al., 1992; Colonna et al., 1990; Petitot et al., 2009c), but no effect on in vivo GI was observed (Casiraghi et al., 1992). However, pasta products were tested by only six volunteers in this study. Heat is known to induce the formation of protein aggregates. Cysteines residues and disulphide bonds that are not accessible in the native conformation can become available and may react to form intermolecular cross-links (Visschers & de Jongh, 2005). Colonna et al. (1990) thus hypothesised that the formation of a high level of protein intermolecular cross-linking induced by VHT drying may lead to more intense starch encapsulation, thus decreasing its susceptibility to enzymes.
The structure of pasta has been well studied in order to identify the factors which control starch degradation. In comparison, the impact of processing on the protein digestibility has received less attention. Yet, the problem of protein degradability in the gastrointestinal tract is of particular relevance for proteins that act as food allergens after ingestion. Allergenicity of a food can be modified by changing the IgE-binding epitopes of the potential allergens. An epitope is the distinctive molecular shape on the allergenic proteins which interacts with antibodies. Linear epitopes are made from a few amino-acids of the polypeptide chain in their linear order while conformational epitopes are made from a few amino acids from different parts of the sequence, brought together by the folding of the polypeptide chain (Davis, 1998). No single figure can be traced to link structure and allergenicity : allergenicity results from a complex balance between formation and destruction of IgE-binding epitopes. Changing 3D-structure of allergens, for example by processes like thermal treatments, can lead to abolition of conformational epitopes and/or masking/demasking of epitopes. Obviously, allergenicity also relates to susceptibility of - 35 -
Chapitre 1 : Synthèse bibliographique
allergens to enzymes as digestion of proteins may lead to destruction of some epitopes. Then, mechanisms that increase resistance to enzymes (such as aggregation) are likely to increase allergenicity. In particular, thermal treatments can induce modifications of the physicochemical and immunological characteristics of the potential allergens by affecting the protein digestibility (Davis, 1998). Only a few studies have been devoted to the characterisation of durum wheat (Triticum durum) allergens compared to bread wheat (Triticum aestivum) but it has been demonstrated that semolina allergens have an allergenic potential similar to that of bread wheat flour (Simonato et al., 2004). Gluten proteins were suggested to be responsible for IgE-mediated reactions (exercise induced anaphylaxis and atopic dermatitis) after ingestion of wheat-based products (Simonato et al., 2004); in particular, gliadins were identified as important allergens (Rasanen, Lehto, Turjanmaa, Savolainen & Reunala, 1994; Varjonen, Vainio & Kalimo, 1997; Varjonen, Vainio & Kalimo, 2000). Since pasta processing modifies protein structure, it may also affect the potential allergenicity of wheat proteins. For example, Simonato, Pasini, Giannattasio, Peruffo, De Lazzari & Curioni (2001) demonstrated that the baking process decreased the digestibility of the potential allergens in bread. These authors compared the protein digestibility and allergenicity of unheated bread dough, bread crumb and bread crust which can be considered as two heat treatments of different intensities (180°C respectively). High temperatures induced protein aggregation that involved probably mainly disulphide bonds and hydrophobic interactions in the bread crumb and strong irreversible interactions (different from hydrophobic interactions or disulphide bonds) in the bread crust (Pasini, Simonato, Giannattasio, Peruffo & Curioni, 2001; Simonato et al., 2001). This protein aggregation prevented the complete proteolytic degradation of the allergenic proteins which could elicit the allergic response in the gastrointestinal tract. In contrast to bread, processing semolina into cooked pasta did not change the total protein digestibility, although cooked pasta was more resistant to pepsin digestion compared to semolina. Moreover, all the potential allergens (> 1 kDa) were degraded after peptic/pancreatic digestion in both semolina and cooked pasta, although the presence of potentially allergenic protein fragments with very low Mr ( < 1 kDa) in digested pasta cannot be excluded (Simonato et al., 2004). However, as seen for bread, it was demonstrated that the application of a very high drying temperature (90-110°C) decreased the in vitro digestibility of proteins in cooked pasta by about 10% compared to a low drying temperature (De Zorzi et al., 2007; Petitot et al., 2009c). The higher resistance of proteins to digestion could be attributed to the presence of highly aggregated proteins stabilised by covalent protein interactions, such as interpeptide cross-linking and Maillard-type aggregates (De Zorzi et al., 2007; Petitot et al., - 36 -
Chapitre 1 : Synthèse bibliographique
2009c). Indeed, nonenzymatic browning related to the Maillard reaction readily occurs during pasta drying, especially at high and very high temperatures (Anese, Nicoli, Massini & Lerici, 1999; Resmini & Pellegrino, 1994). In contrast to what was observed in bread, a modification of the degradability of proteins did not affect their allergenic properties after ingestion (De Zorzi et al., 2007; Petitot et al., 2009c). More studies should be carried out to confirm these results. Moreover, the application of high drying temperatures could induce a decrease in the lysine bioavailability—about 30% of lysine could become unavailable in pasta processed under hightemperature drying cycles (Resmini et al., 1994). This decrease in lysine bioavailability could be due to the protein-carbohydrate Maillard reactions which involve reducing sugars and the terminal amino group of a free amino acid, in particular lysine (Acquistucci, 2000).
4. Conclusion Among cereal products, pasta is an interesting source of slow release carbohydrates. Attempts have been made to correlate the low glycaemic index of pasta with its structural properties at different scales. The compact structure of pasta, the encapsulation of starch by proteins and the physical structure of starch (retrogradation degree, amylose-to-amylopectin ratio) are the main hypotheses that have been put forward to explain the reduced enzymic susceptibility of starch in cooked pasta. This specific structure is the result of successive changes occurring throughout the pasta making process. Forming and drying appear to be essential industrial steps which should be monitored in order to control the carbohydrate digestion of pasta. As the extrusion step gives pasta its compact structure, it would be interesting to determine whether a modification in the extrusion parameters could modify the density of extruded pasta and further decrease the digestibility of starch in cooked pasta. At the molecular level, starch and protein components undergo minor changes during extrusion. The drying step mainly affects the protein fraction. The use of high drying temperatures would allow preservation of the protein network at the microscopic scale, even after prolonged cooking. At the molecular level, the higher the drying temperature, the higher the protein insolubilisation and aggregation. Moreover, depending on the moisture and drying conditions applied, very high temperature may promote the formation of protein aggregates stabilised by irreversible protein interactions, probably through interpeptide cross-linking or Maillard-type reactions. Structural modifications obtained especially at high drying temperatures could strengthen the protein network, and therefore better preserve starch from enzymic attack in cooked pasta. However, very high drying temperatures have been shown to decrease protein - 37 -
Chapitre 1 : Synthèse bibliographique
digestibility and lysine bioavailability (an essential amino acid). Moreover, the use of high drying temperatures would not affect the potential allergenicity of wheat proteins after ingestion. Processes should thus be controlled so as to maintain or enhance the nutritional qualities of food products. This supposes the control of structural transformations of protein and starch fractions but questions remain unanswered. Should processes be optimised/developed in order to change the structural state of the protein or/and the starch fractions? Should starch granules be maintained in their native form (thus preventing starch damage and gelatinisation) or transformed into resistant starch? What is the best balance between the formation of a strengthened protein network (which protects starch from enzymic attack) and the decrease in lysine bioavailability, for example? Regardless of the answer, control of the pasta nutritional properties would require control of the food structure, which depends on the process parameters applied. However, so far only a few authors have assessed, in a single study, the impact of processing on pasta structure as well as the inherent consequences on its nutritional properties. For pasta, as for any food, a multidisciplinary and integrated approach is now needed for rational food design, in order to be able to control digestion and nutrient absorption through the food structure. This implies that food should no longer be regarded as a simple source of essential nutrients but instead as an assembly of macromolecules forming a complex structure resulting from structural modifications occurring at different scales during the manufacturing process. Control of the nutritional properties of a food would require the control, adaptation of standard processes or even development of new processes.
Acknowledgements This work was carried out under the « Programme National de Recherche en Alimentation et nutrition humaine », project « ANR-05-PNRA-019, PASTALEG» that received the financial support of the « ANR- Agence Nationale de la Recherche - The French National Research Agency »
- 38 -
Chapitre 1 : Synthèse bibliographique
Partie 2. Publication 2: Fortification of pasta with legume flour: processing ease, cooking quality, structure and nutritional properties Maud Petitot, Valérie Micard En cours de rédaction, à soumettre
ABSTRACT One of the great challenges today is the development of inexpensive foods which are nutritionally good and at the same time acceptable to the intended consumers. Pasta is a staple food which is known to have a low glycemic index. Despite this interesting nutritional property, pasta is poor in fibres and lacks some essential amino acids, i.e. lysine and threonine. Some work has thus been initiated in order to improve the nutritional composition of pasta. Nutritionally enhanced pasta could be obtained by introducing non traditional ingredients such as legume flours. This review focuses on the nutritional interest of combining wheat and legume in a single food product. It presents also the limiting factors (processing ease, consumer acceptability) that must be taken into account to develop such pasta products.
Key words: Durum wheat pasta, Legume, Nutrition, Cooking quality
- 39 -
Chapitre 1 : Synthèse bibliographique
1. Introduction One of the great challenges today is the development of inexpensive foods which are nutritionally good and at the same time acceptable to the intended consumers. Pasta products meet these specifications. From a nutritional point of view, it is a source of slow release carbohydrates: starch is slowly digested and absorbed in the small intestine, hence promoting a low plasma glucose response (Björck, Liljeberg & Ostman, 2000; Jenkins et al., 1983). Despite this interesting nutritional quality, as most cereal products, pasta is deficient in lysine and threonin (Abdel-Aal & Hucl, 2002; Kies & Fox, 1970) and has a low fibre content. Non traditional ingredients can be introduced in pasta to improve its protein composition (Manthey & Hall, 2007; Nielsen, Sumner & Whalley, 1980; Rayas-Duarte, Mock & Satterlee, 1996; Sadeghi et al., 2008) or increase its fibre content (Brennan & Cleary, 2005; Brennan, Kuri & Tudorica, 2003; Cleary & Brennan, 2006; Edwards, Biliaderis & Dexter, 1995; Ingelbrecht, Moers, Abecassis, Rouau & Delcour, 2001; Manthey & Schorno, 2002) . Among these non traditional ingredients, legumes represent a good source of both proteins and fibres but also of minerals and vitamins essential in human nutrition (Tharanathan & Mahadevamma, 2003). However, the introduction of high level of legume flour into pasta, interesting from a nutritional point of view, may cause technological issues and may change the culinary properties of pasta. The objective of this review was therefore to demonstrate the nutritional interest of combining wheat and legume in a single food product and to determine the impact of such association on the pasta production process and pasta quality.
2. Composition of durum wheat and some grain legumes 2.1. Global composition
Durum wheat (Triticum durum) semolina is recognised as the most suitable raw material for pasta production due to its unique colour, flavour and cooking quality (Feillet et al., 1996). Durum wheat is mainly composed of starch (64-68%), proteins (11-18%), dietary fibres (~8%), and lipids (2%) (Table 1.2). In comparison, grain legumes contain generally lower amounts of starch (30-65%) but higher amounts of proteins (18-41%) and dietary fibres (5-34%) especially insoluble ones. Lupin is different from other legumes as it contains very low amounts of starch (3%) and very high amounts of lipids
- 40 -
Chapitre 1 : Synthèse bibliographique
Table 1.2. Global composition of durum wheat and some grain legumes (% db) Dietary fibres Total Insoluble Soluble
Lipids
Starch
Proteins
Durum wheat (Triticum durum) (4, 5, 11)
64-68
11-18
8.1
6.7
1.4
2
Faba bean (Vicia faba) (1, 2)
30-44
28-33
15-24
nd
nd
1-2
Pea (Pisum sativum) (1, 2, 3, 11)
32-48
22-26
13-22
20.3
1.7
1-3
Chickpea (Cicer arietinum) (3, 7, 9, 10, 11)
43-58
18-27
10-18
6-12
3-5
4 -7
Lentille (Lens culinaris) (3, 7, 9, 10)
43-65
23-27
8-20
5-17
2-3
2-3
Lupin (Lupinus albus) 0-3 37-41 34-39 30-34 4-5 7-10 (1, 6, 8) (1) Colonna & Mercier (1979); (2)Cuq & Leynaud-Rouaud (1992); (3) De Almeida Costa, de Silva Queiroz-Monici, Pissini Machado Reis & Costa de Oliveira (2006); (4) D’egidio, Mariani, Mardi, Novaro & Cubadda (1990) ; (5) El-Khayat, Samaan & Brennan (2003); (6) Guillon & Champ (2002); (7) Jood, Bishnoi & Sharma (1998); (8) Mohamed & Rayasduarte (1995); (9) Perez-hidalgo, GuerraHernandez & Garcia-Villanova (1997); (10) Rehman & Shah (2005); (11) Souci, Fachmann & Kraut (1989). nd : not determined
2.2. Proteins In wheat, gliadins (alcohol-soluble) and glutenins (acide/alkali soluble) represent 80% of total proteins vs. 20% for albumins (water-soluble) and globulins (salt-soluble) (Osborne, 1907) (Table 1.3). Glutenins are composed of high molecular weight (HMW) subunits linked by disulphide bonds (Schofield, 1986). They can form both intra and inter-molecular bonds whereas gliadins can form only intramolecular ones (Kokini, Cocero, Madeka & de Graaf, 1994; Singh et al., 2004). The elastic properties of gluten can be ascribed to the large glutenin polymers and their aggregation state, while the gliadin monomers are responsible for the plastic properties. Legume proteins do not contain gluten protein but a major fraction of globulins and albumins (Table 1.3). In split pea and faba bean flours, globulins represent 45-60% and 55-75% of proteins, respectively, and albumins represent 14-30% and 2-33% of proteins respectively (Gueguen & Barbot, 1988; Koyoro & Powers, 1987; Leterme, Monmart & Baudart, 1990; Pasqualini, Lluch & Antonielli, 1991).
- 41 -
Chapitre 1 : Synthèse bibliographique
Table 1.3. Solubility of wheat and some legume proteins Amount (% total protein)
Solvant
Molecular weight (Da)
- Glutenin (insoluble)
30
none
> 1 000 000
- Glutenin
20
Acidic/ alkali solution
> 100 000
- Gliadin (prolamin)
35
Ethanol 70%
25-75 000
- Globulin
5
NaCl 0.5N
20-90 000
- Albumin
15
Water
5-30 000
- Globulin
55-80
NaCl 0.5N
20 000- 70 000
- Albumin
21-28
Water
23 000- 47 000
- Prolamin
nd
Ethanol 70%
nd
- Glutelin
nd
Acidic/ alkali solution
nd
- Insoluble
14-22
none
- Globulin
55-75
NaCl 0.5N
19 000 - 66 000
- Albumin
2-33
Water
17 000 - 91000
- Prolamin
3
Ethanol 70%
2500-21 800
- Glutelin
18
Acidic/ alkali solution
nd
2-20
none
nd
Proteins Wheat (2)
Pea (3, 4, 5)
Faba bean (1, 6)
- Insoluble
(1) El Fiel, El-Tinay & Elsheikh (2002); (2) Feillet (2000); (3) Gueguen and Barbot (1988); (4) Koyoro & Powers (1987); (5) Leterme, Monmart & Baudart (1990); (6) Pasqualini, Lluch & Antonielli (1991).
2.3. Starch
2.3.1. Starch granule structure
Starch is composed of two polyosides, amylose and amylopectin, both of which consist of glucose as a basic building block. Amylose is a linear polymer mainly linked in the form α(1-4) D-glycopyranosyl whereas the branched amylopectin contains additional α(1-6) linkages (Figure 1.5). Starch is a semicrystalline material, i.e., the granules contain alternating crystalline and amorphous regions (Buleon et al., 1998).
- 42 -
Chapitre 1 : Synthèse bibliographique
Figure 1.5. Chemical structure of amylose and amylopectin
Wheat starch presents a bimodal distribution with large (25-40µm) lenticular starch granules (A-type) and small (5-10µm) spherical starch granules (B-type) (Figure 1.6) and are characterized by a amylose to amylopectin ratio of 25/75 (Buleon et al., 1998). In comparison, legume starches are generally composed of oval or irregularly shaped starch granules (Figure 1.6) with average diameter between 3-40 µm for faba bean and pea starches and 10-22 µm for chickpea starch. Amylose and amylopectin are in the same ratio than in cereal (22-35% amylose) except for wrinkled-seeded pea which contains twice as much amylose (Colonna et al., 1979).
Figure 1.6. Wheat and legume starch granules observed by scanning electron microscopy (Petitot, not published) 2.3.2. Crystallinity Starch is characterised by a semi-crystalline structure with varying polymorphic types and degrees of crystallinity. Native starch granules exhibit three types of X-ray diffraction diagrams: the type A for cereal starches, the type B for tuber and amylose-rich starches, and the type C (a mixture of A and B-types) for legume starches (Buleon et al., 1998).
- 43 -
Chapitre 1 : Synthèse bibliographique
2.3.3. Gelatinisation properties
When heated in excess of water, starch granules absorb water, lose their native crystalline order and swell: this phenomenon is known as gelatinisation. As granules continue to expand, amylose leaches from the intergranular phase to the aqueous phase. Cooling the gelatinized starch/water mixture to room temperature leads to the re-crystallisation of starch molecules (mainly amylose) : this process is know as retrogradation (Parker & Ring, 2001). A schematic representation of structural transformations of starch granules during heating and cooling is shown in Figure 1.7.
Figure 1.7. Schematic representation of hydrothermal transformations of a starch granule
Several analysis techniques are used to study different aspects of the gelatinisation process : differential scanning calorimetry (DSC), birefringence and X-ray diffraction (Baks, Ngene, van Soest, Janssen & Boom, 2007). Gelatinisation is an endothermic transition which is well identified with differential scanning calorimetry (DSC). When viewed under polarised light, starch shows birefringence in the form of the typical Maltese cross. Birefringence reflects the high degree of molecular order. Wide and short angle X-ray diffraction can be used to follow short range order (crystalline double helice) and long-range order (alternating crystalline and amorphous lamellae). Depending on the starch concentration, the determination of the degree of starch gelatinisation can vary depending on the method used. At low starch concentration (10% w/w), DSC, X-ray and birefringence generally give similar degree of gelatinisation. At high concentration (60%, w/w), each method can lead to different results (Baks et al., 2007). Based on DSC measurements, durum wheat starch presents generally lower onset, peak and conclusion temperatures compared to legume starches (Table 1.4).
- 44 -
Chapitre 1 : Synthèse bibliographique
Table 1.4. Thermal characteristic (DSC) of durum wheat and some legume starches: onset (To), peak (Tp) and conclusion (Tc) temperatures (°C) T0
Tp
Tc
(3, 4, 5)
50.9 - 54.4
56.3 - 58.7
63.1 - 63.6
Chickpea
(2)
57.9
63.5
70.4
Yellow pea
(2)
58.2
65.1
70.4
Garbanzo bean
(1)
62.8
68.1
nd
Durum wheat
(1) Czuchajowska, Otto, Paszczynska & Byung-Kee (1998); (2) Huang, Schols, van Soest, Jin, Sulmann & Voragen (2007);(3) Jacobs, Eeringen, Rouseu,Colonna & Delcour (1998);(4)Vansteelandt et al.(1999); (5)Yue et al.(1999).
3. Nutritional value of wheat and grain legumes and their food products 3.1. Nutritional value of wheat and wheat products 3.1.1. Starch Dietary carbohydrates are characterised with respect to both chemical composition and rate and extent of starch digestion. The glycemic index (GI) has been introduced in order to rank foods based on their postprandial blood glucose response (Jenkins et al., 1981). It is based on the measurement of the postprandial glycemic area of a test meal, expressed as the percentage of the corresponding area of a reference food (glucose or white bread) (Figure 1.8). Increasing the consumption of low GI carbohydrates in the diet have been associated with a range of health benefits, including protection against diabetes, coronary heart disease, obesity and even colon cancer (Jenkins et al., 2002; Opperman, Venter, Oosthuizen, Thompson & Vorster, 2004). Among wheat products, pasta is considered as a source of slow-release carbohydrates, therefore possessing a low GI (Table 1.5) (Fosterpowell & Miller, 1995; Granfeldt et al., 1991a; Jenkins et al., 1983). Several hypotheses have been suggested to explain the slow rate of starch digestion in pasta such as the compact structure of pasta (Fardet et al., 1998a; Granfeldt et al., 1991a), the encapsulation of starch granules by fibres (Tudorica et al., 2002) and proteins (Colonna et al., 1990; Fardet et al., 1998b), the structural state of starch granules such as it degree of gelatinisation, retrogradation and amylose-to-amylopectin ratio (Akerberg et al., 1998; Holm & Bjorck, 1988b).
- 45 -
Chapitre 1 : Synthèse bibliographique
Figure 1.8. Schematic representation of glycemic index calculation
Table 1.5. Glycemic index (GI) of wheat products, other cereal products or legumes, with glucose or bread as the standard food GI (Glucose = 100)
GI (Bread = 100)
Wheat product - Corn flakes (Kellogg)
(2)
80 ± 6
114
- Bread (white flour)
(2)
69 ± 5
99
Spaghetti (durum)
(1)
53
76 ± 12
Linguine (durum)
(1)
43
62 ± 11
- Rice, white, low amylose
(3)
88 ± 11
126
- Rice, white, high amylose
(3)
64 ± 9
91
- Maize sweet corn
(2)
59 ± 11
84
- Chickpea
(2)
36 ± 5
51
- Navy bean
(2)
31 ± 6
44
- Pasta
Other cereal products
Cooked grain legumes
Measured GI (in healthy volunteers) appear in bold character while calculated GI appear in italic. When glucose was given as a standard food, the GI against a bread standard was calculated by multiplying the GI values by 1.42 (100/70, GI of bread = 70 when glucose is the standard). When bread was the standard food, the GI value was calculated by multiplying by 0.7 (Fosterpowell et al. (1995)) (1) Granfeldt et al (1991b) ; (2) Jenkins et al. (1981) ; (3) Miller, Pang & Bramall (1992).
- 46 -
Chapitre 1 : Synthèse bibliographique
3.1.2. Proteins
Wheat protein is considered to be of poor quality because it has insufficient amounts of two essential amino acids: lysine, the first-limiting amino acid; and threonine, the second-limiting one (Abdel-Aal et al., 2002; Arrage, Barbeau & Johnson, 1992). Processing semolina into pasta, drying at low temperature (40°C) and cooking do not affect the in vitro protein digestibility (about 87% for durum wheat semolina and pasta) but can result in a reduction in lysine, methionine, tryptophan, and histidine by 8, 16, 13 and 11%, respectively (Abdel-Aal et al., 2002). However, increasing drying temperatures of pasta may lead to reduced in vitro protein digestibility (De Zorzi et al., 2007) but also increased loss of total and available lysine in cooked pasta (Acquistucci, 2000; Anese et al., 1999; Dexter, Tkachuk & Matsuo, 1984; Resmini et al., 1994). This decrease in lysine bioavailability could be due to the protein-carbohydrate Maillard reactions which involve reducing sugars and the terminal amino group of a free amino acid, in particular lysine (Acquistucci, 2000; Anese et al., 1999). Moreover, wheat protein can act as food allergen in sensitised individuals. Durum wheat semolina (Triticum durum) allergens have an allergenic potential similar to that of bread wheat flour (Triticum aestivum) (Simonato et al., 2004). Gluten proteins were suggested to be responsible for IgE-mediated reactions (exercise induced anaphylaxis and atopic dermatitis) after ingestion of wheat-based products (Simonato et al., 2004); in particular, gliadins were identified as important allergens (Rasanen et al., 1994; Varjonen et al., 1997; Varjonen et al., 2000). Allergenicity of a food can be modified by changing the IgE-binding epitopes of the potential allergens. An epitope is the distinctive molecular shape on the allergenic proteins which interacts with antibodies. Linear epitopes are made from a few amino-acids of the polypeptide chain in their linear order while conformational epitopes are made from a few amino acids from different parts of the sequence, brought together by the folding of the polypeptide chain (Davis, 1998). Changing 3D-structure of allergens, for example by processes like thermal treatments, can lead to abolition of conformational epitopes and/or masking/demasking of epitopes. Obviously, allergenicity also relates to susceptibility of allergens to enzymes as digestion of proteins may lead to destruction of some epitopes. Then, mechanisms that increase resistance to enzymes (such as protein aggregation) are likely to increase allergenicity. In particular, thermal treatments can induce modifications of the physicochemical and immunological characteristics of the potential allergens by affecting the protein digestibility (Davis, 1998). According to Simonato et al. (2004), processing semolina into cooked pasta does not change the total protein digestibility. Moreover, all the potential allergens (> 1 kDa) were degraded after peptic/pancreatic digestion in - 47 -
Chapitre 1 : Synthèse bibliographique
both semolina and cooked pasta, although the presence of potentially allergenic protein fragments with very low Mr (< 1 kDa) in digested pasta cannot be excluded Simonato et al. (2004). However, de Zorzi et al. (2007) demonstrated that the application of a ultra high drying temperature (180°C) decreased the in vitro digestibility of proteins in cooked pasta compared to a low drying temperature probably due to large protein aggregates stabilised by strong irreversible interactions formed during drying. In contrast to pasta dried at low temperature, whose potential allergens are degraded during the in vitro digestion process, protein aggregates (with allergenic potential) formed during ultra high temperature drying are resistant to the action of the digestive enzymes. The application of such treatment could thus increase the potential allergenicity compared to pasta dried at low temperature.
3.2. Nutritional value of legumes
Grain legumes would play a key role in the prevention of cardiovascular disease (due to the high fibre content, the low GI and the presence of oligosaccharides), diabetes (due to the low GI and the high content in non-digestible fibres), digestive tract disease (due to the high content in insoluble fibres which accelerate the transit in the intestinal tract), overweight and obesity (due to the satiety effect) (Duranti, 2006). Moreover, legume seeds contain significant amounts of vitamins and minerals (de Almeida Costa et al., 2006; Tharanathan et al., 2003). Despite these interesting nutritional properties, legume seeds contain also a number of antinutritional compounds such as phytic acid and α-amylase inhibitors (Duranti, 2006).
3.2.1. Starch
The slow rate of starch digestion in legume seeds can be explained by several factors (Tharanathan et al., 2003). The presence of rigid cell walls would inhibit starch swelling and dispersion. The crystallinity of starch would also be important: C-type starch found in legume is generally more resistant to digestion. Moreover, legumes contain also high amount of resistant starch (RS), e.g., starch that is not digested in the small intestine but later fermented by the colonic microflora. Three types of resistant starch were identified (Englyst, Kingman & Cummings, 1992): physically inaccessible starch RS1 (whole grain cereal and legumes), native starch granules RS2 (raw foods), and retrograded starch RS3 (in cold spaghetti and legume).
- 48 -
Chapitre 1 : Synthèse bibliographique
3.2.2. Oligosaccharides Legumes contain α-galactosides (oligosaccharides) (Table 1.6) that are not digested in the upper part of the gastrointestinal tract, due to the absence of α-galactosidase among human endogenous enzyme. They are fermented in the colon and induce the production of gas (CO2, H2 et CH4) and short-chain fatty acids (Guillon et al., 2002). The production of gases is responsible for the digestive discomfort related to legume consumption. The production of short-chain fatty acids may have several beneficial implications for health due to the stimulation of the bifidobacteria: potential protective effects against colorectal cancer, improvement of carbohydrate and lipid metabolism, increasing the availability of essential minerals (Guillon et al., 2002). Lupin, pea and ckickpea seeds contain high amounts of α-galactosides compared to faba bean. α-galactosides can be partially or totally removed by soaking, or germination (Table 1.6) (Guillon et al., 2002; Torres, Frias, Granito & Vidal-Valverde, 2007b).
Table 1.6. α-galactosides content of some grain legumes (% db) α-galactosides Lentil (1)
3.0-7.1
Chickpea (1)
7.4 -7.5
Faba bean (1)
3.1- 4.2
Lupin (1)
7.4 -9.5
Pea (1)
5.1- 8.7
Pigeon pea (2)
5.5
Germinated pigeon pea (2)
1.0
(1) Guillon et al. (2002) ; (2) Torres et al.(2007b).
3.2.3. Proteins Legumes are an important source of food proteins and are considered as “poor man’s meat”. Legume proteins are relatively low in sulphur-containing amino acids, methionine, cystein and tryptophan but the amounts of lysine are much greater than in cereals grains (Alonso, Orue, Zabalza, Grant & Marzo, 2000; Carbonaro, Cappelloni, Nicoli, Lucarini & Carnovale, 1997) (Table 1.7). Wheat and legume present amino-acid complementarities.
- 49 -
Chapitre 1 : Synthèse bibliographique
Table 1.7. Amino-acid scores of durum wheat semolina and grain legumes, based on FAO/WHO/UNU (1985) pattern for a pre-schooled child (2-5 years)
Essential amino-acids
FAO pattern (mg/g proteins) (4)
Durum wheat
Faba bean
Pea
Chickpea
Lupin
Lentil
(1)
(3)
(2)
(3)
(5)
(3)
Histidine
19
111
141
116
135
174
139
Isoleucine
28
111
159
112
162
134
160
Leucine
66
109
123
126
118
118
123
Lysine
58
38
117
142
117
84
118
25
156
74
86
102
100
75
63
125
127
146
142
89
119
Threonine
34
91
118
114
109
103
119
Tryptophan
11
100
nd
85
nd
55
nd
Valine
35
114
146
156
137
117
151
Methionine + Cysteine Phenylalanine + Tyrosine
(1) Abdel-Aal & Hucl (2002); (2) Alonso et al, (2000); (3) Carbonaro et al. (1997); (4) FAO/WHO/UNU (1985); (5) Sujak, Kotlarz & Strobel (2006). nd: not determined
Despite their interesting nutritional composition legume seed proteins can act as food allergens. Microwave heating, boiling, and extrusion cooking produce minimum changes on the IgE recognition of lupin allergens (Alvarez-Alvarez et al., 2005). Conversely, boiling (100 °C, 15 min) of lentil seeds produces a strong increase in allergenicity as demonstrated by ELISA inhibition experiments (Sandin, San Ireneo, Lizana, Fernandez-Caldas, Lebrero & Borrego, 1999). Therefore, thermal-processing procedures at various temperatures and conditions could have unpredictable effects on the allergenic activity of legume proteins. They can create epitopes as well as destroying existing ones (Wal, 2003). In pasta, one article reported an allergic reaction in lupin fortified pasta (Hefle, Lemanske & Bush, 1994).
- 50 -
Chapitre 1 : Synthèse bibliographique
4. Fortification of pasta with legume flour The most common legume flours used to fortify pasta are chickpea (Goni & ValentinGamazo, 2003; Sabanis, Makri & Doxastakis, 2006; Wood, 2009; Zhao, Manthey, Chang, Hou & Yuan, 2005), lupin (Lampart-Szczapa, Obuchowski, Czaczyk, Pastuszewska & Buraczewska, 1997; Rayas-Duarte et al., 1996; Torres, Frias, Granito, Guerra & Vidal-Valverde, 2007a), pea (Frias, Kovacs, Sotomayor, Hedley & Vidal-Valverde, 1997; Guillon et al., 2002; Nielsen et al., 1980; Torres et al., 2007b; Zhao et al., 2005) and lentil (Bahnassey & Khan, 1986a; Zhao et al., 2005). These studies mainly focused on the impact of legume flour addition on the culinary properties of pasta. Less attention has been devoted to the study of structural and nutritional properties of legume pasta.
4.1. Technological issues
According to most authors, the introduction of legume flour to pasta did not required essential changes in the production process (Rayas-Duarte et al., 1996; Torres et al., 2007a; Zhao et al., 2005). However, in most studies, the addition of legume flour did not exceed 30% (w/w) and pasta was produced at a laboratory-scale. However, higher levels of legume fortification and/or the production of pasta at a pilot or industrial-scale could induce technological issues. Lorenz (1979) reported that 40% appeared to be the maximum possible replacement level of faba bean flour. At higher levels, it became rather difficult to extruded pasta properly with a pasta press. Wood (2009) reported that fortification of pasta with chickpea flour (10-30%) made the dough particles increasingly sticky, causing them to aggregate during mixing, making the extrusion step difficult. This author suggested that a lower water addition could be used compared to the durum wheat dough in order to improve mixing and extrudability of chickpea fortified dough. In accordance with these results, Sabanis et al. (2006) studied the mixing properties of wheat-chickpea flour with a farinograph and found that the amount of water required to develop a dough of 500 farinograph units at the peak of the curve increased with chickpea flour addition. They also studied the stretching properties of the dough with the Brabender extensograph. Up to 30% level substitution, the dough is still strong and elastic but at upper level (30-50%), the dough is weakened as a result of the dilution of the gluten.
- 51 -
Chapitre 1 : Synthèse bibliographique
4.2. Cooking and organoleptic properties of legume fortified pasta
Pasta fortified with lupin flour presented good textural and organoleptic properties up to about 10-15% (w/w) of legume addition (Lampart-Szczapa et al., 1997; Rayas-Duarte et al., 1996; Torres et al., 2007a). Higher levels of substitution induced an increase in pasta firmness, adhesiveness and grittiness (Rayas-Duarte et al., 1996) without affecting the flavour (LampartSzczapa et al., 1997; Rayas-Duarte et al., 1996). Higher cooking loss could also be observed in lupin fortified pasta (Rayas-Duarte et al., 1996; Torres et al., 2007a). The ability to form a gluten matrix is unique to wheat flour and semolina and is believed to be the main factor in forming the internal spaghetti network that holds the pasta together. The addition of nongluten flours diluted the gluten strength and interrupted and weakened the overall structure of the spaghetti. This may allow leaching of more solids from the spaghetti to the cooking water (Rayas-Duarte et al., 1996). In chickpea fortified pasta, the overall acceptability was reduced at 5-10% (w/w) substitution level (Sabanis et al., 2006; Zhao et al., 2005). In contrast to lupin fortified pasta, the addition of only 5% of chickpea flour induced the development of a nutty and pulse flavour and a raw pulse aroma (Zhao et al., 2005). According to Sabanis et al. (2006), a 5% substitution level gives a different fine taste and made the best impression upon the panellists. From 20% substitution level, the flavour score was reduced. Depending on the authors, the fortification of pasta with chickpea induced high or lower firmness and cooking loss (Wood, 2009; Zhao et al., 2005). Pasta fortified with lentil flour 30% were characterised by higher cooking loss and higher instrumental firmness compared to durum wheat pasta. Sensorial analyses revealed that the overall quality of pasta was reduced at 5% substitution level, probably due to the detection of a raw pulse aroma and nutty flavour by the panellists (Zhao et al., 2005). Increased firmness and stickiness and reduced elasticity were also detected by the panellists at 10, 30 and 20% substitution levels, respectively. In conclusion, up to a 5-10% substitution level, legume fortified pasta are generally well accepted by the panellists but at higher level of substitution, negative changes in the cooking quality (higher cooking loss and stickiness) and sensorial attributes (poor acceptability) appear.
- 52 -
Chapitre 1 : Synthèse bibliographique
4.3. Nutritional properties of legume fortified pasta
The primary objective of legume flour addition to pasta was to increase its nutritional composition. Less attention has been devoted to the study of pasta nutritional properties such as its glycemic index or its protein digestibility.
4.3.1. Nutritional composition
Fortification of pasta with 10 and 15% of legume (navy bean, pinto bean or lentil flour) is a way to increase the protein, ash and fibre content of pasta. Moreover, legume fortified pasta shows a better balance for the sulphur amino acids as well as the lysine content (Bahnassey, Khan & Harrold, 1986b). Fortification of pasta with 10% lupin or pigeon pea flour was shown to increase the amount of protein, dietary fibre, mineral, vitamin B1, B2 and E contents and antioxidant capacity compared to wheat pasta (Torres et al., 2007a; Torres et al., 2007b). Legumes contain antinutritional factors such as trypsin inhibitors and lectins (hemagglutinins). Trypsin inhibitors inhibit the action of trypsin in the digestive tract of humans. They could reduce protein nutritional quality and may lead to pancreatic hypertrophy (Zhao et al., 2005). Hemagglutinins agglutinate red blood cells and are very toxic for the intestinal cells. Legumes are usually cooked at high temperature which inactivates trypsin inhibitors and hemagglutinins. However, care should be taken if legume are used in products treated at low temperatures (Bahnassey et al., 1986b). Chickpea flour contains higher trypsin inhibitor followed by lentil, yellow pea and green pea flours. The trypsin inhibitors of spaghetti containing 15% of green pea or lentil flour are totally eliminated after cooking, but in pasta containing 15% chickpea flour, 30% of the original trypsin inhibitors remain after cooking (Zhao et al., 2005).
4.3.2. Glycemic index (GI)
The starch digestibility of spaghetti fortified with 25% chickpea flour was investigated by Goni et al. (2003). In vitro measurements revealed that starch was digested more slowly in chickpea fortified spaghetti than in wheat spaghetti These results were confirmed in vivo: the GI of chickpea fortified pasta (58 ± 6) was significantly lower than the GI of wheat spaghetti (73 ± 5). These authors suggested that this could be explained by a higher amount of indigestible fraction (non-starch polysaccharides, resistant starch, oligosaccharides) in chickpea pasta. In accordance with these results, Osorio-Diaz, Agama-Acevedo, Mendoza-Vinalay, Tovar & Bello- 53 -
Chapitre 1 : Synthèse bibliographique
Perrez (2008) found that the in vitro starch hydrolysis of pasta fortified with 40% chickpea flour was reduced from 84 to 62% compared to durum wheat pasta. They attributed this lower starch digestibility to the higher content in resistant starch and resistant starch associated fibres (starch remnants in dietary fibre residues) of chickpea flour.
4.3.3. In vitro protein digestibility
The fortification of pasta with 5-30% of lupin flour was shown to increase significantly the in vitro protein digestibility of cooked pasta (from 85.5% for wheat pasta to 86.3% for lupin fortified pasta) (Rayas-Duarte et al., 1996). The supplementation of pasta with 10% αgalactoside free lupin flour did not change the in vivo protein digestibility in rats (Torres et al., 2007a). However, the protein efficiency ratio (PER) increased from 1.11 for wheat pasta to 1.92 for lupin fortified pasta. The PER has been used extensively for predicting the protein quality of foods. It is calculated as the amount of weight gain (g) per unit of protein consumed. Fortification of pasta with 20% cowpea flour did not have a significant effect on the in vitro digestibility of proteins (~79%) (Nur Herken, Ibanog lu, Oener & Ibanog Iu, 2006).
4.3.4 Future trends
The development of nutritionally enhanced wheat products and especially pasta products should be investigated in the future. First results have shown that the addition of legume flour to pasta increases its nutritional composition. It may also affect its starch and protein digestibility but phenomenon involved remain unclear. The study of structural transformations induced by the incorporation of legume flour to pasta could give some answers. Indeed, food structure is recognized of prime importance to control the nutritional properties of foodstuffs. In particular, the low glycemic index (GI) of pasta is generally ascribed to its specific compact structure (Bjorck, Granfeldt, Liljeberg, Tovar & Asp, 1994; Wolever et al., 1986) and/or the presence of a strong protein network entrapping starch granules (Colonna, Leloup & Buleon, 1992; Fardet et al., 1998b). A change in the nature and proportion of components (higher amount of fibres, dilution of the gluten network) in legume pasta may affect its structure and therefore its GI. However, to the best of our knowledge, the impact of legume flour addition on pasta structure has not been investigated yet. The study of mixing properties of wheat-chickpea flour blends (cf. 3.1) by Sabanis et al. (2006) revealed that the dough containing over 30% of chickpea flour was
- 54 -
Chapitre 1 : Synthèse bibliographique
weakened as a result of the dilution of the gluten. This shows that legume flour may potentially affect pasta structure and that further investigations are thus needed.
5. Conclusion Wheat and legumes are nutritionally complementary. Their combination in a single food product such as pasta is a way to benefit from the nutritional value of both raw materials (richness in proteins, fibres, slow release carbohydrates, essential amino-acids). In addition to an improved nutritional composition, the addition of legume flour to pasta could further decrease its starch digestibility. However, the level of legume flour addition has to be controlled because above 10-15% supplementation with legume flour, the quality (texture and flavour) of pasta is decreased. Moreover, high level of legume flour addition may induce the formation of a sticky dough during mixing, and prevent its extrusion. The development of legume fortified pasta requires therefore a compromise between processing ease, consumer acceptability and nutritional gain.
- 55 -
- 56 -
Chapitre 2 Matériels et méthodes
- 57 -
- 58 -
Chapitre 2 : Matériels et Méthodes
Chapitre 2. Matériels et méthodes
1. Caractérisation des matières premières La semoule de blé dur (Triticum durum) de qualité supérieure a été fournie par Panzani (campagne 2005-2006). Les farines de fève (Vicia faba) et de pois cassé (Pisum sativum) dépelliculées ont été fournies par Terrena (campagne 2005-2006).
1.1. Teneur en eau des matières premières (publications 3, 4, 5 et 6)
Un échantillon est pesé dans une coupelle de masse connue, placé dans une étuve à 130°C pendant 2h puis dans un dessiccateur pendant 20min. La coupelle est de nouveau pesée et la perte en eau est ensuite exprimée en g d’eau absorbée pour 100 g d’échantillon. Teneur en eau de la farine (g d' eau/100g de farine) =
Mh - Ms Mh
Matière sèche de la farine (g de MS / 100g de farine) = 100 - teneur en eau
Mh: masse initiale (humide) de l’échantillon de farine (g) Ms : masse sèche de l’échantillon de farine (g)
1.2. Composition chimique des matières premières (publications 4, 5 et 6)
- Amidon : la teneur en amidon total des matières premières est déterminée au moyen d’un kit enzymatique (Megazyme, Co., Wicklow, Irlande). Le principe repose sur la solubilisation de l’amidon par cuisson en présence d’α-amylase thermostable. Les dextrines obtenues sont ensuite hydrolysées en glucose par une amyloglucosidase. Le glucose est alors dosé par spectrophotométrie à 510 nm après incubation avec un mélange de glucose oxydase et de peroxydase.
- 59 -
Chapitre 2 : Matériels et Méthodes
- Protéines : un dosage de l’azote total est effectué par la méthode de Kjehldal. Le facteur de conversion de l’azote utilisé pour les protéines de légumineuses (qui tient compte de la richesse des protéines en azote) est de 5,7. Les mesures sont réalisées en triplicata. L’échantillon (contenant de l’azote organique) est minéralisé par l’acide sulfurique, en présence d’un catalyseur, puis après alcalinisation, l’ammoniac libéré (azote minéral) est distillé, recueilli dans une solution d’acide borique, et enfin dosé par colorimétrie (formation d'un complexe coloré bleu-vert entre l'ion ammonium, le salicylate de sodium et le chlore en milieu alcalin). L'absorbance du complexe formé (catalysé par du nitroprussiate de sodium) est mesurée à 630 nm (cette méthode ne distingue pas l’azote protéique de l’azote non-protéique).
- Fibres : elles ont été déterminées en duplicata par l’ISHA (Lonjumeau, France) selon la méthode AOAC 985.29.
- Lipides : les analyses des lipides ont été réalisées par l’un des partenaires du projet (UMR 476 INSERM / 1260 INRA, Marseille, France). La technique d’extraction des lipides utilisée est basée sur la méthode de Bligh and Dyer (1959). Trois extractions sont réalisées pour chaque farine de légumineuse et six pour la semoule de blé dur.
- Cendres : elles sont analysées après incinération à 900°C selon une norme française (NF 03720).
1.3. Granulométrie des matières premières et de l’amidon (publications 4 et 5)
Elle est déterminée au moyen d’un granulomètre laser en voie liquide (type Beckman Coulter LS 230, Fullerton, USA). Le principe repose sur la diffraction d’un faisceau laser au contact de particules. L’angle de déviation du faisceau dépend de la taille de ces dernières.
1.4. Solubilité des protéines et coagulation à la chaleur (publication 5)
Un échantillon de 4g de semoule de blé dur, de farine de fève ou de pois cassé est mis en suspension dans 36 ml d’eau désionisée puis agité 15min à 30 tours/min (position 2 modèle Heidolph Reax2, Milan, Italie). La suspension est centrifugée 15min à 12000g (centrifugeuse type Beckman Avanti centrifuge TM J-30 I, Fullerton, USA). Un dosage des protéines solubles est effectué selon la méthode de Bradford, sur les surnageants obtenus. Cette technique est basée sur une réaction colorimétrique entre les protéines et le Bleu de Coomassie G-250 de couleur rouge à l’état libre. L’ajout de ce réactif à une solution de protéines entraîne la formation d’un complexe coloré bleu dont le maximum d’absorbance est mesuré à 595 nm. Le contenu protéique du culot - 60 -
Chapitre 2 : Matériels et Méthodes
a été déterminé par la méthode de Kjeldahl pour calculer la solubilité des protéines. Un aliquote (1,5 ml) du surnageant obtenu précédemment est placé dans un tube Eppendorf de 2 ml et plongé dans de l’eau à 100 ± 2°C pendant 4 min puis immédiatement mis dans la glace. L’échantillon est ensuite centrifugé 15min à 13000g (Centifugeuse type Sigma 3K10, Osterode, Allemange). Le dosage de protéines dans le surnageant selon la méthode de Badford avant et après cuisson permet de déterminer le pourcentage de protéines coagulées. Deux répétitions sont effectuées pour chaque matière première.
1.5. Propriétés d’hydratation des matières premières (publication 4)
Elles sont évaluées au moyen du Plastographe (Brabender OGH, Duisburg, Allemange) munie d’une cuve Farinographe (Brabender OGH, Duisburg, Allemagne) reliée à un système d’hydratation contrôlé au moyen d’une seringue en verre (Fortuna Optima Glasspritze, Poulten & Graf) reliée à un système de pompe (KD Scientific, model kdS 100, USA). 50 g de poudre sont introduits dans la cuve munie de deux pâles tournants en sens inverse. Le système est raccordé à la seringue dont le débit est de 1ml/min. Le couple résistant sur les pâles est mesuré en fonction du temps. Le pourcentage d’eau minimal nécessaire au développement de la pâte est déterminé.
2. Fabrication des pâtes alimentaires Les pâtes alimentaires sont fabriquées à l’aide d’une presse pilote (Bassano, Lyon, France). Trois étapes fondamentales constituent le procédé de fabrication : hydratation/malaxage, extrusion, et séchage (Figure 2.1).
2.1. Fabrication des pâtes alimentaires au blé dur (publications 3, 4, 5)
- Hydratation/ malaxage : 5 kg de semoule sont hydratés avec de l’eau du robinet (environ 1.3 L) au moyen d’une rampe d’hydratation placée dans le malaxeur, de façon à obtenir une teneur en eau de 47% (bs). Le mélange est malaxé pendant 5 min à 120 tr/min (phase d’hydratation) puis pendant 15 min à 60 tr/min (phase de malaxage) en changeant de sens de rotation toutes les minutes.
- Extrusion : après avoir fait le vide, la semoule hydratée est extrudée à travers un moule à spaghetti, à une température contrôlée de 35°C. Lorsque la pression d’extrusion est stable
- 61 -
Chapitre 2 : Matériels et Méthodes
(environ 110 bars), les échantillons de pâtes alimentaires sont collectés à la sortie de la filière. Les brins de spaghetti sont coupés à l'aide de ciseaux en morceaux de 50-60 cm de long et disposés sur des cannes métalliques.
Figure 2.1. Présentation de la presse pilote et de ses accessoires
- 62 -
Chapitre 2 : Matériels et Méthodes
- Séchage : les pâtes alimentaires fraîches sont séchées dans un séchoir pilote (AFREM, Lyon, France) à humidité relative (HR) et température contrôlées (88%, 35°C). Quatre programmes de séchage sont appliqués : séchage basse température (55°C) (LT), séchage haute température (70°C) (HT), séchage très haute température (90°C) avec application de la très haute température en début de cycle (VHT) ou en fin de cycle (VHT.LM) (Figure 2.2). À la sortie du séchage, les pâtes alimentaires sont stockées pendant une semaine dans des sacs en plastique, avant caractérisation.
Figure 2.2. Diagrammes de séchage appliqués au cours du séchage des pâtes dans le séchoir pilote. HR : humidité relative ; T : temperature
- 63 -
Chapitre 2 : Matériels et Méthodes
2.2. Fabrication des pâtes enrichies en légumineuse (publications 4, 5, 6) - Hydratation/ malaxage : 5 kg d’un mélange semoule de blé dur/ farine de légumineuse (65/35) sont mélangés seuls pendant 5 min à 120 tr/min (phase d’homogénéisation) puis avec de l’eau pendant 15 min à 120 rpm (phases d’hydratation et de malaxage) de façon à obtenir une teneur en eau de 45% (bs), en changeant de sens de rotation toutes les minutes. - Extrusion : les pâtes sont ensuite extrudées comme pour les pâtes au blé dur. - Précuisson et/ou séchage : après extrusion, les pâtes fraîches sont séchées dans le séchoir pilote ou séchées par lyophilisation ou précuites avant d’être séchées. Les pâtes séchées dans le séchoir pilote sont séchées selon le programme LT, VHT ou VHT.LM. Les pâtes séchées par lyophilisation sont conditionnées dans des sachets plastique au congélateur (-20°C) immédiatement après extrusion puis lyophilisées à l’échelle pilote par la société Lyofal (Salon de Provence, France) une semaine après leur fabrication. Les pâtes précuites sont précuites dans une autoclave horizontale (120°C/1.2 bars / 10 min) (B.B.C, Montpellier, France) avant d’être séchées dans le séchoir pilote selon le programme LT.
2.3. Propriétés d’agglomération des matières premières au cours du malaxage (publication 4) Afin de déterminer la machinabilité de la semoule et des mélanges semoule/farine de légumineuse (65/35), leurs propriétés d’agglomération sont évaluées en fin de malaxage par une méthode de tamisage mesurant la distribution de taille des agglomérats de la masse sableuse. En fin de malaxage, un aliquote de la masse sableuse est transféré sur une colonne de tamis de taille décroissante (de 4 et 1 mm) afin de classer les petites particules (d < 1 mm), les agglomérats (1 < d < 4 mm) et les morceaux de pâte (4 mm < d). Les tamis sont agités manuellement pendant 30 sec pour éliminer les agglomérats obstruant la grille. Chaque tamis est ensuite pesé. La distribution en taille est exprimée en pourcentage de la masse totale sur trois répétitions.
3. Tests de cuisson 3.1. Temps optimal de cuisson (publications 3, 4 ,5 ,6) La cuisson des pâtes est réalisée selon un protocole standardisé (NF ISO 7304). Un échantillon de 100 g de pâtes découpées en brins de 15 cm environ est plongé dans 2l d'eau Evian bouillante contenant 7 g/L de sel. À intervalles de temps réguliers, un brin de spaghetti est - 64 -
Chapitre 2 : Matériels et Méthodes
prélevé dans l'eau de cuisson puis immédiatement écrasé entre deux plaques de plexiglas, afin de visualiser la ligne blanche correspondant à l’amidon non gélatinisé (Figure 2.3). Le temps optimal de cuisson (OCT), correspond au temps à partir duquel la ligne blanche a totalement disparue (amidon complètement gélatinisé). La cuisson des pâtes est arrêtée à OCT + 1 min pour toutes les analyses sur pâtes cuites. Les pâtes sont alors égouttées 30 secondes à l’aide d’une passoire.
Figure 2.3. Illustration du test de la ligne blanche utilisé pour la détermination du temps optimal de cuisson. A : pâte non cuite, B : pâte cuite, C : pâte cuite au temps optimal de cuisson
3.2. Cinétique d’absorption d’eau au cours de la cuisson (publications 5 et 6)
Les cinétiques d’absorption d’eau des pâtes au cours de la cuisson sont suivies selon la méthode de del Nobile, Baiano, Conte & Mocci (2005). Des brins de spaghetti de 40 ± 0.5 mm sont découpés, pesés puis plongés dans des tubes (un brin par tube) contenant 9 ml d’eau distillée, équilibrés dans un bain thermostaté à 100°C ± 1°C. Les brins de spaghetti sont enlevés à des intervalles de temps réguliers (toutes les 30 secondes pendant les 10 premières minutes puis toutes les minutes jusqu’à 15min), égouttés sur un papier filtre puis de nouveau pesés. Le ratio entre le poids gagné et le poids initial du spaghetti sec (∆W/W0)= W(t) – W0/W0 est calculé à chaque temps. Trois essais sont réalisés pour chaque type de pâte.
- 65 -
Chapitre 2 : Matériels et Méthodes
3.3. Mesure de l’eau absorbée dans les pâtes cuites (OCT + 1min) (publication 4)
L’eau absorbée est évaluée sur un échantillon de 50g de pâte sèche et calculé comme suit :
Eau absorbée (% ms) =
Masse pâtes cuites (OCT + 1 min) − Masse pâtes sèches × 100 Masse pâtes sèches
3.4. Pertes à la cuisson (publication 4)
La matière sèche (MS) des pâtes sèches est tout d’abord déterminée. Une quantité connue de pâtes est cuite jusqu’au temps OCT + 1min. La matière sèche des pâtes cuites est déterminée après un séchage en deux étapes. La première étape consiste à sécher les pâtes à 50°C pendant 2 jours. Les pâtes peuvent alors être broyées, un aliquote est prélevé et séché à 130°C pendant 2 heures. Les pertes à la cuisson sont ensuite calculées.
Pertes à la cuisson (%, ms) =
MS pâtes cuites (g ) − MS pâtes sèches (g ) × 100 MS pâtes sèches (g )
4. Caractérisation des principaux constituants des pâtes 4.1. Caractérisation de l’amidon
4.1.1. Analyse enthalpique différentielle (publications 3, 5, 6)
Les mesures d’analyse enthalpique différentielle (AED) (ou differential scanning calorimetry (DSC)) sont conduites au moyen d’une DSC 2019 modulée (publication 3) ou d’une DSC Q200 modulée (publications 5et 6) (TA Instruments, New Castle, USA), calibrée à l’indium. Les analyses sont effectuées sur les pâtes sèches et les pâtes cuites (ayant été congelées à l’azote puis lyophilisées) broyées puis tamisées de manière à obtenir une poudre de taille inférieure à 250 µm. L’échantillon est pesé dans une capsule hermétique en aluminium. De l’eau distillée est ajoutée à l’aide d’une micropipette de manière à avoir un ratio eau/solide de 4. La capsule est ensuite sertie, pesée et mise à équilibrer pendant 1h à température ambiante. Cette capsule et une capsule de référence vide sont mises en place dans deux compartiments identiques du four de l’appareil. Les capsules sont ensuite chauffées entre 10°C et 120°C avec une vitesse
- 66 -
Chapitre 2 : Matériels et Méthodes
de chauffe de 10°C/min. La différence d’énergie calorifique instantanée à fournir entre les deux capsules pour les maintenir à la même température est mesurée à chaque instant au cours du chauffage. Un exemple typique de thermogramme d’un amidon de blé chauffé en excès d’eau est présenté Figure 2.4. Chaque endotherme est caractérisée par : la température de début de transition (To), la température de fin de transition (Tf), la plage de température dans laquelle l’événement thermique est détecté (∆ Tr = Tf – To), la température au sommet du pic (Tp) et la variation d’enthalpie associée à cette transition (∆H).
Figure 2.4. Exemple de thermogramme obtenu par analyse enthalpique différentielle (DSC)
4.1.2. Microscopie électronique à balayage laser environnemental (MEBE)(publication5)
Les granules d’amidon extraits de la semoule de blé dur et des farines de légumineuses sont observés par microscopie électronique à balayage laser environnemental (MEBE) ou electronical scanning electron microscopy (ESEM)
Extraction - Semoule de blé dur : 100g de semoule sont mixés avec de l’eau distillée (1 :1) au moyen d’un mixer (KM221, Kenwood, France) jusqu’à ce que la pâte commence à coller aux parois, indiquant le début du développement du gluten. De l’eau est alors ajoutée en grande quantité (200ml) et mélangée ensuite pendant 1min de manière à extraire l’amidon de la matrice de gluten. La suspension est filtrée au moyen d’un tamiseur en voie liquide (Retsch AS 200 digit,
- 67 -
Chapitre 2 : Matériels et Méthodes
GmbH & Co, Germany), muni de différents tamis (200, 160, 120, 75 and 50 µm). Le filtrat obtenu est centrifugé (3000g/20min/4°C) (Beckman Avanti centrifuge TM J-30 I, Fullerton, USA) et la couche supérieure jaune-marron contenant la fraction d’amidon riche en protéines est séparée de la couche inférieure contenant l’amidon à extraire. L’amidon est purifié par différentes étapes de lavage à l’eau et centrifugations (3000g/20min/4°C) jusqu’à disparition complète de la couche supérieure jaune-marron. L’amidon purifié est séché sous la hotte à température ambiante pendant une nuit. - Amidon de légumineuse : 100g de farine de pois cassé ou de fève sont mélangés dans de l’eau distillée (1 : 6) ajustée à pH 9 avec 1M NaOH de manière à augmenter la solubilité des protéines (Carbonaro et al., 1997). La suspension est filtrée, le filtrat est centrifugé puis l’amidon est ensuite purifié comme précédemment.
Observation Un microscope électronique à balayage laser environnemental (Xl 30 ESEM, Philips, Netherlands) équipé d’un détecteur secondaire au gaz est utilisé pour l’observation des granules d’amidon extraits des matières premières. L’échantillon d’amidon est placé sur un support puis observé à température ambiante avec une tension de 20 kV et une pression de 2 Torr. Pour compléter les observations microscopiques, une analyse de distribution en taille des granules d’amidon est réalisée au moyen d’un granulométre laser en voie liquide (type Beckman Coulter LS 230, Fullerton, USA). Le principe repose sur la diffraction d’un faisceau laser au contact de particules. L’angle de déviation du faisceau dépend de la taille de ces dernières.
4.2. Caractérisation de l’état d’agrégation des protéines (publications 3, 5 et 6) L’état d’agrégation des protéines a été suivi par chromatographie liquide haute performance d’exclusion stérique (size-exclusion high performance liquiq chromatography ou SE-HPLC) après extraction des protéines sur la semoule de blé dur, les mélanges semoule de blé dur/farine de légumineuse (65/35), les poudres en fin de malaxage, les pâtes extrudées, les pâtes sèches et les pâtes cuites. Les échantillons de poudre en fin de malaxage, de pâtes extrudées et de pâtes cuites ont été préalablement congelés à l’azote, lyophilisés puis broyés.
- 68 -
Chapitre 2 : Matériels et Méthodes
4.2.1. Extraction des protéines
L’extraction des protéines est conduite selon la méthode de Morel, Dehlon, Autran, Leygue & Bar-L'Helgouac'h (2000) sur 160 mg d’échantillon lyophilisé et broyé. Les protéines sont extraites après deux extractions successives. Une première extraction est conduite à 60°C pendant 80 min, sous agitation rotative, dans 20ml de tampon phosphate 0,1M pH 6.9 contenant 1% de SDS. Les extraits de protéines sont ensuite centrifugés à 39000g (20°C/30min). 1ml de surnageant (fraction SDS-soluble) est ensuite prélevé pour injection en SE-HLPC. Les protéines insolubles sont extraites du culot par une seconde extraction à 60°C pendant 60 minutes dans 5ml du tampon phosphate SDS contenant 20mM de dithioerythritol (DTE), puis sonifié pendant 5 min pour extraire la fraction des protéines insolubles dans le SDS. La fraction de protéines restant insolubles après les deux extractions constitue la fraction de protéines insolubles.
4.2.2. Distribution en taille des protéines
La distribution en taille des protéines est analysée par chromatographie liquide haute performance (ou SE-HPLC). L’appareil est équipé d’une colonne TSK G4000-SW (Merck, France) (7.5 × 300 mm) et d’une pré-colonne TSK G3000-SW (Merck, France) (7.5 × 75 mm), comme décrit dans Morel et al. (2000). Une fois corrigées des différents ratios solide/solvant durant l’extraction, les aires (en unités arbitraires) des fractions SDS-soluble et DTE-soluble sont ajoutées et la somme est exprimée en pourcentage de l’aire correspondante calculée pour la matière première utilisée pour fabriquer les pâtes (semoule de blé dur dans le cas des pâtes au blé dur et un mélange de 35% de farine de légumineuse + 65% de semoule de blé dur pour les pâtes mixtes). Chaque profil SE-HPLC des extraits SDS-solubles est divisé en 5 fractions majeures (S1 à S5). Les masses moléculaires apparentes sont estimées en calibrant la colonne avec des protéines standards selon Redl, Morel, Bonicel, Vergnes & Guilbert (1999). La fraction S1 correspond aux protéines polymériques éluées dans le volume mort de la colonne (Bleu dextran, MM = 2000 kDa). La fraction S2 correspond aux protéines comprises entre 780 à 95 kDa. Les fractions S3 et S4 correspondent aux protéines comprises entre 95 et 52 KDa et entre 52 et 21 kDa respectivement. La fraction F5 correspond aux protéines monomériques les plus petites ( 50 nm) et mezopores (2-50 nm) est obtenue à partir de la courbe du volume d’intrusion cumulé en fonction de la taille des pores.
5.2. Couleur des pâtes sèches et des pâtes cuites (publication 4) La couleur des pâtes sèches et des pâtes cuites est mesurée à l’aide du chromamètre Minolta (Modèle CR-400, Minolta Co., Osaka, Japon) en utilisant le système Hunter L*, a*, b*. L* représente la luminance allant du noir (valeur : 0) au blanc (valeur : 100). La composante a* représente la gamme allant du rouge (128) au vert (-128). La composante b* représente la gamme allant du jaune (128) au bleu (-128).
- 70 -
Chapitre 2 : Matériels et Méthodes
5.3. Rhéologie des pâtes cuites (publications 3 et 4) Les propriétés rhéologiques des pâtes alimentaires sont déterminées à l'aide d'un texturomètre (TAXTplus, Stable Micro Systems) afin de réaliser deux types de test : un test en extension et un test en compression de type « Texture Profile Analysis » (TPA). Immédiatement après cuisson (OCT + 1 min) et égouttage, les spaghettis sont conservés environ 30 min dans une boite de Pétri en atmosphère saturée d'eau avant analyse.
Test en extension : ce test consiste à étirer un brin de spaghetti à vitesse constante jusqu’à sa rupture. Un brin de spaghetti est fixé au moyen de deux pinces (spaghetti tensile grips A/SPR, Stable Micro Systems). Les paramètres du test sont les suivants : force de détection du produit de 5 g ; distance initiale entre les deux pinces de 15 mm ; vitesse de déplacement du module de 3 mm/s. La courbe de la contrainte en fonction de la déformation est alors tracée : l’aire sous la courbe correspond à l’énergie à la rupture (Figure 2.5).
Figure 2.5. Représentation du test en extension
Test TPA (Texture Profile Analysis) : ce test consiste à appliquer une déformation de 30% sur un morceau de spaghetti à deux reprises au moyen d'un module carré (Stable Micro Systems). On relève la force mesurée au cours du temps. Ce test permet de décrire la texture du spaghetti à l’aide de multiples paramètres. La fermeté (hardness) correspond à la force maximale relevée lors de la première compression. La cohésion (cohesiveness) correspond à l’aire du 2ème pic sur
- 71 -
Chapitre 2 : Matériels et Méthodes
l’aide du 1er pic. La résilience (resilience) représente l’aire de la seconde moitié du 1er pic (aire 2) divisée par l’aire de la première moitié du 1er pic (aire 1). L’élasticité (springiness) représente la distance de la première moitié du 2ème pic (d2) sur la distance de la première moitié du 1er pic (d1) (Figure 2.6). Les paramètres du test utilisés sont les suivants : force de détection du produit de 5 g ; déplacement du module jusqu’à 70% de l’épaisseur initiale du spaghetti (ce qui correspond à une déformation de 30% ((L-L0) / L0) avec L0 le diamètre initial du spaghetti et L la diamètre après compression); vitesse de déplacement du module de 1 mm/s ; temps de repos de 10 s entre les 2 compressions.
Figure 2.6. Représentation du test TPA
5.4. Microsctructure des pâtes cuites (publications 3, 5 et 6)
5.4.1. Préparation des coupes
Après cuisson au temps OCT + 1 min, 3 morceaux de pâtes (3 essais) sont congelés pour chaque échantillon. Un morceau de pâte (de 0.5cm de long) est découpé au moyen d’une lame de rasoir puis déposé sur un support à – 40°C sur lequel une goutte d’eau a préalablement été déposée. Pendant la congélation de l’échantillon, un cryoprotecteur (OCT, Cellpath, Newtown, Royaume-Unis) est ajouté. Les coupes transversales de pâtes sont réalisées au moyen d’un cryotome (Microm HM 560, Walldorf, Allemagne). Des coupes de 8 µm sont réalisées pour la microscopie optique en lumière blanche, en fluorescence et en lumière polarisée. Des coupes de - 72 -
Chapitre 2 : Matériels et Méthodes
15 µm sont réalisées pour la microscopie confocale à balayage laser. Les coupes à congélation sont déposées délicatement sur des lames polysine (CLM, Nemours, France) puis stockées à température ambiante jusqu’à analyse.
5.4.2. Microscopie optique en lumière blanche (publication 5)
Les protéines sont colorées 10 min avec du Fast Green 1g.L-1 (Sigma Aldrich Co., USA) et les granules d’amidon sont colorés 1min avec de l’iode par application d’une solution de Lugol (Fluka, Buchs, Suisse), diluée au 1/8 (v/v). Les coupes sont ensuite rincées à l’eau distillée (3 fois/ 2 min). Les protéines apparaissent ainsi colorées en vert et les granules d’amidon en bleumarron. Les images sont ensuite observées au moyen d’un microscope Leica DM6000M équipé d’une caméra Leica DFC290 digitale et du logiciel Leica Qwin image (version 3.0) (Leica Microsystems, Wetzlar, Allemagne). 5.4.3. Microscopie optique en lumière polarisée (publications 5 et 6)
La perte de biréfringence, témoignant de la gélatinisation de l’amidon, est suivie sous lumière polarisée. Les coupes non colorées sont observées au moyen d’un microscope Olympus BX61 (Olympus America Inc., Center Valley, USA), équipé d’une caméra Hamamastu ORCAAG (Hamamatsu Photonics K.K, Hamamatsu, Japon) et du logiciel Cell^P (Olympus America Inc., Center Valley, USA). 5.4.4. Microscopie optique en fluorescence (publication 5)
Les protéines sont tout d’abord éteintes par une coloration de 10 min au Fast Green 1g.L-1. Les fibres sont ensuite colorées 10 min avec du Calofluor 0.1g.L-1 White M2R préparé dans un tampon phosphate de sodium 0.2 mol L-1 (pH 8.0). Les images sont observées avec un microscope en epi-fluorescence Olympus BX61 (Olympus America Inc., Center Valley, USA), équipé d’une caméra Hamamastu ORCA-AG (Hamamatsu Photonics K.K, Hamamatsu, Japon) et du logiciel Cell^P (Olympus America Inc., Center Valley, USA). Le Calcofluor est excité par un filtre BandPass à 400-410 nm et détecté en utilisant un filtre LongPass à 455 nm.
- 73 -
Chapitre 2 : Matériels et Méthodes
5.4.5. Microscopie confocale à balayage laser (ou CLSM) (publications 3, 5 et 6)
Le microscope confocale à balayage laser est utilisé pour visualiser le réseau protéique des pâtes cuites. Une analyse d’image permet ensuite de caractériser la distribution du réseau protéique au sein des pâtes.
- Coloration : les protéines sont colorées 10 min à l’acide fuchsine 0.01% (m/v) préparé dans 1% (v/v) d’acide acétique. Les coupes sont ensuite rincées à l’eau distillée (3 fois / 2 min).
- Acquisition des images : les images sont acquises en niveaux de gris entre 0 et 255 avec un microscope confocal à balayage laser inversé (Zeiss Axiovert 200M) muni du logiciel Zeiss LSM510 META (Carl Zeiss, Jena, Allemagne). 3 brins de spaghetti sont coupés pour chaque type de pâte et sur chaque brin de spaghetti, 3 coupes sont observées. Pour chaque coupe transversale, 3 images sont enregistrées, une image dans la zone externe, une image dans la zone intermédiaire et une image dans la zone centrale.
- Analyse d’image : une filtration par transformée de Fourier (FFT) est tout d’abord réalisée afin de réduire le bruit du signal. Une procédure de normalisation des niveaux de gris est ensuite conduite afin de corriger les différences de niveaux de gris dues à l’hétérogénéité des coupes. Les images observées sont divisées par une image calculée de « blanc ». L’image de blanc est obtenue par dilatation de l’image à l’aide d’un élément structurant de taille 5 puis en utilisant la procédure de fermeture d’aire (Guillemin, 2003). Une fois la procédure de normalisation terminée, le réseau protéique est caractérisé par morphologie mathématique (Devaux, Robert, Melcion & deMonredon, 1997; Rouille, Della Valle, Devaux, Marion & Dubreil, 2005). Cette méthode consiste à transformer l’image de manière irréversible selon le niveau de gris de ses pixels. Un masque, l’élément structurant, de forme carré, est déplacé sur l’image. Le niveau de gris du pixel central est modifié selon les niveaux de gris de ses voisins recouverts par l’élément structurant. L’opération de base érosion attribue au pixel central la valeur minimale en niveau de gris observé sur la zone recouverte par l’élément structurant : l’image s’assombrit. L’opération inverse, la dilatation, lui attribue la valeur maximale : l’image s’éclaircit. Un exemple de courbe d’érosion dilatation obtenue sur des pâtes cuites est présentée Figure 2.7. La partie droite de la courbe correspond à l’érosion et caractérise le film protéique du plus fin au plus épais. La partie gauche de la courbe correspond à la dilatation et caractérise les cellules noires pouvant être assimilées aux granules d’amidon, des plus petits aux plus gros. La somme des niveaux de gris
- 74 -
Chapitre 2 : Matériels et Méthodes
(Volume V) des images est calculée après chaque étape d’érosion/dilatation et les courbes granulométriques ou courbes d’érosion dilatation sont obtenues par l’équation :
g (i) =
V(i) − V(i + 1) V0 − Vf
V0 représente le volume initial de l’image, Vf le volume après la dernière étape d’érosion/dilatation et i une des étapes d’érosion/dilatation. Les courbes g mesurent la proportion du volume qui a été modifié entre l’étape i et (i+1), c’est à dire entre deux érosions ou dilatations successives. Les courbes d’érosion dilatation sont ensuite analysées par analyse en composant principale (ACP). Une analyse de variance (ANOVA) est ensuite réalisée sur les scores des composantes principales. L’analyse d’image est réalisée au moyen du logiciel Matlab (v7.0.4) (The MathWorks, Paris, France) en utilisant des boîtes à outils dédiées (image processing and PLS toolbox v3.5 (Eigenvector Research Inc., Manson USA)).
Figure 2.7. Exemple de courbe d’érosion-dilatation obtenue après analyse d’image des pâtes au blé dur cuites.
- 75 -
Chapitre 2 : Matériels et Méthodes
5.5. Analyse sensorielle des pâtes (publication 4)
Les analyses sensorielles descriptives sont réalisées sur pâtes cuites (OCT +1min) par un panel composé de 14 juges entraînés du Centre Européen de Recherche sur les Céréales, le Riz et les Pâtes alimentaires de Panzani (CRECERPAL, Marseille, France). Les échantillons sont présentés à chaque juge séparément dans un ordre randomisé. Une note sur une échelle de 1 à 10 est attribuée pour chaque caractéristique ; 1 représentant une intensité faible et 10 une intensité élevée. Les caractéristiques suivantes sont évaluées : - Aspect lisse: propriété de surface indiquant l’absence de craquelure ou de rugosité (1 : rugueux, 10 : lisse). - Homogénéité: différence de texture en bouche entre l’extérieur et l’intérieur du spaghetti (1: hétérogène; 10: homogène). - Fermeté: force nécessaire pour atteindre une pénétration donnée (NF ISO 5492, 1992). (1: mou; 5: ferme, 10: dur). - Elasticité : capacité de la pâte déformée à revenir à son état initial après que la déformation ait cessé (NF ISO 5492, 1992) (1: plastique; 5: élastique; 10: caoutchouteux). - Friabilité : force nécessaire pour que la pâte s’effrite ou se brise. C’est le résultat d’une fermeté élevée et d’un faible degré de cohésion (NF ISO 5492, 1992) (1: friable; 5: croquant; 10: cassant).
5.6. Propriétés nutritionnelles (publications 3, 5 et 6)
5.6.1. Digestibilité in vitro de l’amidon par mesure du glucose rapidement disponible (RAG) (publications 5 et 6)
Les analyses de digestibilité in vitro de l’amidon des pâtes cuites sont réalisées par Englyst Carbohydrates Ltd. (Englyst et al, 1999). La méthode est basée sur la mesure par HPLC de la quantité de glucose libérée durant un temps d’incubation donné avec des enzymes digestives, dans des conditions standardisées. Le glucose libéré après 20 min d’incubation correspond au glucose rapidement disponible (ou rapidly available glucose : RAG). Il est exprimé en g pour 100 g de glucides disponibles (c’est à dire les glucides totaux moins l’amidon résistant). Le RAG mesuré in vitro est un bon prédicteur de la glycémie postprandiale (Englyst et al, 1999).
- 76 -
Chapitre 2 : Matériels et Méthodes
5.6.2. Digestibilité in vitro de l’amidon et des protéines (publication 3) Les analyses de digestibilité in vitro de l’amidon et des protéines dans les pâtes au blé dur cuites ont été réalisées au sein de l’unité de recherche Biopolymères, Interactions et Assemblages (BIA, INRA, Nantes). Les pâtes sont broyées avec un broyeur à viande de manière à reproduire la mastication buccale (Hoebler, Deavux, Karinthi, Belleville & Barry, 2000). Une digestion in vitro en batch, est ensuite conduite en trois phases : une phase buccale en présence d’amylase, une phase gastrique en présence de pepsine à pH 2 et une phase intestinale en présence de pancréatine à pH 7. L’analyse de la digestibilité de l’amidon et des protéines est ensuite conduite sur les digestats à différents moments de la phase bucco-gastrique et de la phase intestinale. Le degré d’hydrolyse de l’amidon est déterminé par mesure de l’amidon transformé en dextrines solubles dans l’éthanol 80% (ayant un degré de polymérisation compris entre 1 et 10-12). La concentration en équivalent-glucose dans les extraits éthanoliques et les échantillons est évaluée par la méthode au phénol-sulfurique (Dubois, Gilles, Hamilton, Rebers & Smith, 1956). Le degré d’hydrolyse de l’amidon est ensuite calculé :
DH amidon (%) =
Glucides extractibles × 100 Glucides totaux
Le degré d’hydrolyse des protéines est déterminé par l’augmentation du nombre de fonctions amines primaires (NH2) dans les extraits protéiques (Frister, Meisel & Schlimme, 1988). Le nombre d’amines primaires est déterminé par un dosage o-phtaldialdéhyde (OPA). Le degré d’hydrolyse est calculé à partir du nombre d’amines primaires initialement présents (T0), libérés pendant la digestion (Tx) et obtenus après une hydrolyse totale des protéines (Ttotal).
DH protéines (%) =
[ NH 2 ] ( Tx ) − [ NH 2 ] ( T 0) [ NH 2 ] ( Ttotal ) − [ NH 2 ] ( T 0)
× 100
5.6.3. Allergénicité (publication 3)
Le potentiel allergène des protéines a été évalué au sein de l’unité de recherche Biopolymères, Interactions et Assemblages (BIA, INRA, Nantes) par dot-blots sur les digestats à l’aide de sérums de patients allergiques au blé ou aux graminacées. Afin de connaître l’origine protéique de certains fragments ayant résisté à la digestion, des ELISA compétitifs ont également été réalisés avec les jus de digestion et un pool de sérums de patients allergiques au blé. - 77 -
- 78 -
Chapitre 3 Pâtes alimentaires au blé dur Influence des barèmes de séchage sur leur structure et leurs propriétés nutritionnelles
- 79 -
- 80 -
Chapitre 3. Pâtes alimentaires au blé dur
Chapitre 3. Pâtes alimentaires au blé dur. Influence des barèmes de séchage sur leur structure et leurs propriétés nutritionnelles
Les travaux présentés dans ce chapitre ont été réalisés en collaboration avec l’Unité de recherche Biopolymères, Interactions et Assemblages (BIA, Nantes).
Chapitre 3 : INTRODUCTION Afin d’approfondir les connaissances sur le lien structure – nutrition, une étude a été menée sur les pâtes au blé dur, reconnues comme un aliment à faible index glycémique (IG). Selon les auteurs, ce faible IG peut être attribué notamment à la structure compacte de la pâte et au réseau protéique enchâssant les granules d’amidon, les protégeant ainsi de l’attaque amylasique. Cette structure spécifique de la pâte est le résultat de transformations structurales de l’amidon et des protéines au cours du procédé de pastification. Toute modification de cette structure est donc susceptible de modifier les propriétés nutritionnelles des pâtes et notamment leur IG. Les séchages à haute température, utilisés afin d’améliorer la qualité culinaire des pâtes, peuvent être à l’origine de ces modifications structurales. Certaines études ont en effet montré qu’ils pouvaient modifier l’état structural de l’amidon, l’état d’agrégation des protéines et même les propriétés nutritionnelles des pâtes. Cependant, peu de travaux ont intégré dans une seule étude l’effet des traitements thermiques sur la structure des pâtes et les conséquences sur leurs propriétés nutritionnelles. Ce chapitre est consacré à l’étude de la structure (à différentes échelles) et des propriétés nutritionnelles (digestibilité de l’amidon, digestibilité protéique et potentiel allergène) des pâtes au blé dur soumises à différents traitements de séchage.
- 81 -
Chapitre 3. Pâtes alimentaires au blé dur
- 82 -
Chapitre 3. Pâtes alimentaires au blé dur
Publication 3: Modification of pasta structure induced by high drying temperatures. Effects on the in vitro digestibility of protein and starch fractions and the potential allergenicity of protein hydrolysates
Maud Petitot, Chantal Brossard, Cécile Barron, Colette Larré, Marie-Hélène Morel, Valérie Micard
Food Chemistry 116 (2009) 401-412
ABSTRACT The effects of drying on pasta structure, starch and protein digestibility and potential allergenicity were investigated. Pasta was dried at low (55°C, LT), high (70°C, HT), and very high temperature (90°C) applied either at the beginning (VHT) or at the end of the drying profile (VHT.LM). Changes in dried and in cooked pasta structures were determined regarding protein solubility, thermal properties of starch, microscopic and rheological measurements. Changes were moderate up to 70°C and increased at higher temperatures and especially for VHT.LM drying. VHT.LM drying tended to decrease starch digestibility and significantly decreased protein digestibility of cooked pasta by 10 % probably due to the formation of highly aggregated proteins linked by very strong covalent bonds. None of the drying profile was found to abolish the allergenic properties of pasta. IgE-reactive peptides from prolamins and albumins/globulins fractions were found in all digestion juices, VHT-LM drying increased the latter.
Keywords: Durum wheat pasta; Process; Microstructure; Rheology; SE-HPLC; Nutritional properties
- 83 -
Chapitre 3. Pâtes alimentaires au blé dur
1. Introduction Pasta is a traditional and highly popular cereal-based food product because of its convenience, nutritional quality, and palatability (Cubadda et al., 2007). Pasta is obtained after kneading semolina and water, extruding and drying. The most noticeable recent innovation in pasta production is the application of high temperature in the drying process (Aktan et al., 1992; Cubadda et al., 2007; Degidio, Mariani & Novaro, 1993; Guler et al., 2002; Novaro, d' Egidio, Mariani & Nardi, 1993; Zweifel et al., 2003). This innovation was found to affect positively pasta qualities (higher firmness, lower cooking loss and lower stickiness) (Baiano et al., 2006; de Stefanis & Sgrulletta, 1990; Zweifel et al., 2003), especially when the high temperature is applied during the final stages of the drying process. Up to now, of all the major quality factors, nutritional properties of pasta have received the least attention by researchers. Among cereal products pasta appears to possess unique nutritional features in that the starch is slowly digested and absorbed in the small intestine, hence promoting a low plasma glucose response (Anese et al., 1999; Björck et al., 2000; Jenkins et al., 1983). The low Glycemic Index (GI) of pasta is generally attributed to its compact structure (Barkeling, Granfeldt, Björck & Rossner, 1995; Bjorck et al., 1994; Granfeldt et al., 1991a; Wolever et al., 1986), but other numerous factors have been suggested to explain the different rates of starch degradation. Indeed, the surface area accessible to digestive enzymes (Granfeldt et al., 1991a), the encapsulation of starch granules by fibres (Brennan, Blake, Ellis & Schofield, 1996) and proteins (Colonna et al., 1990; Fardet et al., 1998b), and the physical structure of starch, such as its degree of gelatinisation or retrogradation, and its amylose-to-amylopectin ratio (Akerberg et al., 1998; Holm et al., 1988a) were demonstrated to affect starch digestion. Some work has been initiated to study the effect of process parameters on the nutritional quality of pasta. For example, high drying temperatures were shown to decrease the in vitro digestibility of starch (Casiraghi et al., 1992) and proteins (De Zorzi et al., 2007). Although some hypothesis have been suggested to explain the relationship between pasta structure and its nutritional properties, there is still limited data on the subject. The objective of this work was to better characterize, by a multiscale approach, the structure of pasta dried with different drying profiles, and to study its repercussion on the in vitro digestibility of carbohydrate and protein fractions, and the potential allergenicity of digested pasta.
- 84 -
Chapitre 3. Pâtes alimentaires au blé dur
2. Materials and Methods 2.1. Pasta manufacturing
2.1.1. Pasta production
Spaghetti was processed with a continuous pilot-scale pasta extruder (Bassano, Lyon, France). Durum wheat semolina (Panzani, Marseille, France) (5kg) was hydrated with tap water to obtain a moisture content of 47g/100g db and then mixed for 15min at 60rpm. The product was then extruded at 31rpm and 40°C. Extruded spaghetti was then dried in a pilot-scale drier (AFREM, Lyon, France) in order to reach 12% of moisture. Four drying profiles were applied: Low Temperature 55°C (LT), High Temperature 70°C (HT), Very High Temperature 90°C applied either at the beginning of the cycle, when the moisture content of pasta is high (about 20%) (VHT) or at the end of the drying cycle, when the moisture content of pasta is low (about 12%) (VHT.LM). The relative humidity and temperature inside the dryer and the moisture content of pasta (evaluated according to AACC method (44-15)) were monitored during drying.
2.1.2. Pasta cooking
Dried spaghetti was cooked in Evian water containing 0.7% (w/v) of sodium chloride with a water-to-solid ratio of 20. Optimal cooking time (OCT) was indicated when the white core of the pasta disappeared when squeezed between two glass plates (Approved method 66-50 AACC 2000). All analyses on cooked pasta were made on pasta cooked at OCT+1min (i.e. 10.3 min).
2.2. DSC measurements on dried and cooked spaghetti
Differential Scanning Calorimetry (DSC) measurements were conducted on a DSC 2019 modulated (TA Instruments, New Castle, USA) calibrated with indium and an empty pan as a reference. Durum wheat semolina, dried pasta and freeze-dried cooked pasta samples were ground and sieved to pass through 250µm mesh screen. Ground samples were accurately weighed in aluminium hermetic pans. Water was added with a micropipette with a water-to-solid ratio of 4. Then the pans were sealed, reweighed and allowed to stand for 1h at room temperature. The analyses were performed from 10 to 120°C at a heating rate of 10°C/min using an empty pan as a reference. For each endotherm, the onset (To), peak (Tp) and conclusion (Tc) - 85 -
Chapitre 3. Pâtes alimentaires au blé dur
temperatures, and the gelatinization enthalpy (∆H) were computed by using the TA instruments analysis software program. Temperature ranges (∆Tr = Tc-To) were calculated. Each experiment was repeated four times for each sample. Data were subjected to analysis of variance (ANOVA) followed by the Fisher’s least significant difference (LSD) test to compare means at the 5% significance level by using Microsoft Xlstat software 2008 (Addinsoft, Paris, France).
2.3. Protein extractability of dried and cooked spaghetti
2.3.1. Protein extraction procedure
Proteins were extracted in triplicate from semolina, dried pasta and freeze-dried cooked pasta, according to a modified method of Morel, Dehlon, Autran, Leygue Bar-L’Helgouach (2000). The first extraction was conducted at 60°C for 80min with a sodium phosphate buffer containing 1% of sodium dodecyl sulphate (SDS, 0.1M) and a solid-to-solvent ratio of 8mg/ml in order to extract SDS-soluble proteins. SDS disrupts the electrostatic, hydrophobic and hydrophilic interactions occurring between proteins. After centrifugation, the pellet was suspended at 60°C for 60min with 5ml of the SDS-phosphate buffer containing 20mM of dithioerythritol (DTE) and sonicated for 5min in order to extract SDS-insoluble proteins, referred as DTE-soluble proteins. The combination of sonication and DTE causes the degradation of disulphides bonds that connect the glutenin subunits together (Singh et al., 1990). The remaining pellet made of unextractable proteins represents proteins linked by covalent linkages that were not affected by sonication and DTE (e.g.: isopeptides bonds). Once corrected for their different solid-to-solvent ratios during extractions, areas (in arbitrary units) of SDS-soluble and DTE-soluble proteins were added and the sum (i.e. total extractable proteins) was expressed as percents of the corresponding area calculated for semolina (on equivalent dry protein basis). Data were subjected to analysis of variance (ANOVA) followed by the Fisher’s least significant difference (LSD) test to compare means at the 5% significance level by using Microsoft Xlstat software 2008 (Addinsoft, Paris, France).
2.3.2. Size distribution measurement
The size distribution of proteins in cooked pasta was studied by Size-Exclusion High Performance Liquid Chromatography (SE-HPLC). The SE-HPLC apparatus (Waters model LC Module1 plus) was equipped with an analytical column TSK G4000-SW (Merck, France) (7.5 × - 86 -
Chapitre 3. Pâtes alimentaires au blé dur
300 mm) and a guard column TSK G3000-SW (Merck, France) (7.5 × 75 mm), as previously described (Morel et al., 2000). Each SE-HPLC profile of SDS-soluble and DTE-soluble proteins was arbitrary divided into five peaks (S1 to S5 and P1 to P5 respectively). Apparent molecular weights were estimated by calibrating the column with protein standards according to Redl, Morel, Bonicel, Vergnes Guilbert (1999). Fraction S1 corresponded to polymeric proteins eluted at the void volume of the column (blue dextran, Mr = 2000 kDa). Fraction S2 corresponded to proteins ranging from Mr ≈ 780 to 95 kDa. Fractions S3 and S4 corresponded to proteins ranging from Mr ≈ 95 to 52 kDa and from 52 to 21 kDa respectively. Fraction S5 corresponded to the smallest monomeric proteins (Mr < 21 kDa). The second extract, obtained after solubilisation by the combined action of DTE and sonication, allowed the characterisation of SDS-insoluble proteins whose molecular weight exceeded 2000 kDa before sonication and solubilisation in DTE. Fraction P1 corresponded to Mr > 2000 kDa. Fraction P2 corresponded to protein ranging from Mr ≈ 780 to 116 kDa. Fractions P3 and P4 corresponded to proteins ranging from Mr ≈ 116 to 62 kDa and from 62 to 21 kDa respectively. P5 corresponded to proteins of Mr < 21 kDa.
2.4. Microscopic characterisation of the protein network in cooked spaghetti
2.4.1. Sample preparation
After cooking (OCT+1), water was decanted and pasta was cooled for 15min before sectioning in a covered container at 25°C. Pasta was cut in small pieces and frozen quickly at 40°C in water plus a cryoprotector (OCT, Cellpath, Newtown, UK). Samples were cut at -20°C using a microtome (Microm HM 560, Walldorf, Germany), to obtain 15µm thick transverse sections. After drying they were stained for 10min in a 0.01% (w/v) fuchsin acid solution diluted in acetic acid 1% (v/v) as described by Fardet et al. (1998a). Sections were then rinsed 3 times for 2min in distilled water.
2.4.2. Image acquisition
Images were acquired using an inverted Confocal Laser Scanning Microscope (CLSM) (Zeiss Axiovert 200M) with attached Zeiss LSM510 META imaging system (Carl Zeiss, Jena, Germany) in the Montpellier RIO Imaging (MRI) facility. The excitation wavelength was 543nm and the light emitted over 560 nm was selected by a long pass filter. An x20 lens coupled to a numeric zoom of 1.1 allowed to acquire 1024x1024 pixels images large of 410µm. The pixel size - 87 -
Chapitre 3. Pâtes alimentaires au blé dur
was therefore equal to 0.4µm. Grey levels were coded on 8 bits, giving an intensity scale from 0 (black) to 255 (white). For each transverse section, images were acquired at 3 locations within the pasta: external, intermediate and central zones, without any overlapping between the images. For each sample of cooked pasta dried with one of the four drying profiles, 3 strands were cut and 3 sections per strand were observed. Therefore 108 images were collected and used for subsequent mathematical treatment.
2.4.3. Image processing
Images were pre-processed in order to normalise grey levels over the surface and between images. A FFT filtering was performed before a normalisation procedure considering a shading over the image as already made on CSLM images of tomato cells (Guillemin, Devaux & Guillon, 2004). The protein network was characterized by analysing the grey level granulometry based on mathematical morphology (Devaux et al., 1997; Rouille et al., 2005), with squared structuring elements. The sum of the grey-level values (the volume V) in the images was calculated after each erosion/dilation step and the erosion/dilation curves were obtained according to the equation : g (i) =
V(i) − V(i + 1) V0 − Vf
where V0 represents the volume of the original image, Vf the volume after the last erosion/dilation step and i one of the steps of the image processing. The g curves measure the proportions of the grey-level volume that were modified between steps i and (i+1), i.e. between two successive erosions or dilations. Erosion/dilation curves were analysed by principal component analysis (PCA). Similarity map could be drawn from principal component scores, whereas the loading analysis revealed the weight of each erosion or dilation step in their computation. All the image analysis and data treatments were performed with Matlab (v7.0.4) software (The MathWorks, Paris, France) using dedicated toolboxes (image processing and PLS toolbox v3.5 (Eigenvector Research Inc., Manson USA).
- 88 -
Chapitre 3. Pâtes alimentaires au blé dur
2.5. Rheological properties of cooked spaghetti
2.5.1. Sample preparation
A TA-XTplus (Stable Micro Systems, Scarsdale, USA) texture profile analyser equipped with a windows version of Texture Expert software package (Stable Micro Systems, Scarsdale, USA) was used to evaluate textural properties of cooked spaghetti. After cooking at (OCT+1), pasta was allowed to equilibrate at ambient temperature for 20min in a covered container at 25°C before texture analysis. The variables were recorded through five measurements for each pasta, cooked in two different occasions, totalling to 10 measurements/pasta. Data were subjected to analysis of variance (ANOVA) followed by the Fisher’s least significant difference (LSD) test to compare means at the 5% significance level by using Microsoft Xlstat software 2008 (Addinsoft, Paris, France).
2.5.2. Texture Profile Analysis
The TA-XTplus was equipped with a 35mm cylindrical probe (ref. P/35, Stable Micro Sytems). The probe compressed a single strand of spaghetti at a constant rate of deformation (1mm/s) to 70% of the initial spaghetti thickness. The probe was retracted and held stationary 10s before performing a second compression to 70% of the original spaghetti thickness. From Texture Profile Analysis (TPA) curve, textural parameters of hardness, cohesiveness, resilience and springiness were obtained (Epstein, Morris & Huber, 2002). Spaghetti hardness was defined as the maximal peak force attained during the first compression. Cohesiveness was calculated as the ratio of the area under the second peak to the area under the first peak. Resilience was defined as the ratio of the area under the second half of the first peak to the area under the first half of the same peak. Springiness was calculated as the ratio between the distance of the first half of the second peak to the distance of the first half of the first peak.
2.5.3. Tensile test
The TA-XTplus was equipped with tensile grips (ref. A/SPR, Stable Micro Sytems). The initial distance between the two tensile grips was 15mm. The test was performed at a constant rate of deformation (3mm/s). Breaking stress (N.m-2) and breaking strain (%) were recorded
- 89 -
Chapitre 3. Pâtes alimentaires au blé dur
from the stress-strain curve. The energy (J.m-3) stored in the sample until fracture which corresponds to the area under the stress-strain curve was also calculated.
2.6. In vitro digestion of cooked and minced spaghetti
2.6.1. Enzymes
Porcine pancreatic alpha-amylase (EC 3.2.1.1) from Fluka Biochemika (10080, 1217678, 43.6U/mg), pepsin (EC 3.4.23.1) from Sigma (P7000, porcine stomach mucosa, 61K0197, 596U/mg) and pancreatin (EC 232.468.9) from Sigma (P7545, 045K0673, 8 USP) were used. Enzymes were prepared in phosphate buffer 5mM, pH6.9, NaCl 1M, CaCl2 4mM. Pepsin solution and supernatants of α-amylase and pancreatin solutions (centrifuged 10 min, 1500 g, 10°C) were stored at 4°C until use (i.e. at most 3hours).
2.6.2. In vitro digestion procedure
Cooked pasta (OCT+1) was ground in a meat mincer (Zyliss Tornado, Switzerland) to reproduce buccal mastication (Hoebler, Devaux, Karinthi, Belleville & Barry, 2000). Minced pasta (15g) was mixed 15min with 25mL of phosphate buffer (5mM, pH 6.9, NaCl 1M, CaCl2 4mM) at 37°C before being digested for 5min at 37°C with porcine pancreatic α-amylase (200 U/g wheat starch). pH was adjusted to 2 with HCl 1N and pepsin was added (73,400 U/g wheat protein). Gastric digestion lasted 0 (inactivated pepsin), 30 or 180 min and was stopped by addition of KOH 2N to raise pH to 7. Intestinal digestion by pancreatin (1g/g wheat starch) lasted 0 (inactivated pancreatin), 10, 30 or 180min and was stopped by placing the reactor for 5min in boiling water. Cooled down digests were homogenized (Kinematica AG, PT3000, Switzerland) and stored at - 20°C.
2.6.3. Determination of carbohydrate hydrolysis during digestion
Hydrolysis of starch during digestion was studied by the percentage of starch transformed into alcohol-soluble dextrins (PASD); dextrins with a degree of polymerization between 1 and 10-12 are soluble in 80% ethanol. Equivalent glucose concentration in ethanolic extracts and samples were determined by the phenol-sulfuric method (Dubois, Gilles, Hamilton, Rebers & Smith, 1956). - 90 -
Chapitre 3. Pâtes alimentaires au blé dur
2.6.4. Study of protein hydrolysis during digestion
Digestion extracts were prepared from 1.5g homogenized digests in 10mL tetraborate buffer (0.1M, pH 9.3, 1% (w/v) SDS, 5% (v/v) 2-mercapto-ethanol) gently stirred for 16h at ambient temperature and centrifuged (5,000g for 10min at 10°C then 10,000g for 20min at 10°C). The number of free α and ε NH2 function in amino acids, peptides and proteins was determined on digestion extracts as previously described (Frister, Meisel & Schlimme, 1988) at the initial phase of digestion (T0), during digestion (Tx) and after a total hydrolysis of T0 (HCl 6N, 24 h at 105°C) (Ttotal). The degree of hydrolysis (DH) was calculated according to the equation:
DH(%) =
[ NH 2 ]( Tx ) − [ NH 2 ]( T 0) [ NH 2 ]( Ttotal ) − [ NH 2 ]( T 0)
× 100
As low remaining protease activity was previously reported for pig pancreatic α-amylase contrary to human salivary α-amylase Fardet et al. (1998b), T0 corresponding to the end of the buccal phase was chosen as the reference for protein hydrolysis. The soluble peptides released during the digestion were analyzed by gel filtration ; homogenized digests were centrifuged twice (first at 1,500g for 10 min and second at 10,000g for 10min at 20°C), to prepare the so called digestion juices. High performance liquid chromatography was carried out with a Alliance 2795 HPLC System (Waters S.A.S., Saint-Quentin En Yveline, France) onto a Superdex™ Peptide 10/300 GL with a fractionation range of Mr 100 to 7000 and 13µm average particle size. Digestion juice was mixed (v/v) with trifluoroacetic acid (TFA) 0.2%, filtered on Millipore 45µm and samples of 100µL were loaded onto the column which was eluted with acetonitrile 30%-TFA 0.1% at 0.5ml/min for 60min. Dual detection at 220 and 280nm was performed.
2.7. IgE dot blotting and competitive ELISA
Sera were obtained with informed consent from 4 patients with food allergy to wheat and 1 patient allergic to Graminaceae without food allergy (control serum). All patients with allergy to wheat were children and suffered from atopic eczema/dermatitis syndrome associated for two of them with asthma, and one had digestive symptoms upon pasta consumption. The serum from the last patient was used for IgE dot blotting experiment. A pool of the four sera was used for competitive ELISA. - 91 -
Chapitre 3. Pâtes alimentaires au blé dur
For IgE dot blotting, 2µl of digestion extracts or digestion juices from pasta and four controls (gliadin fraction 0.5mg/ml, purified α-gliadin 0.5mg/ml, albumin/globulin fraction 0.2 mg/ml and a mixture of α-amylase, pepsin and pancreatin, each at concentration corresponding to the digestion assay) were applied on nitrocellulose sheet (0.2µm, Sartorius, Germany) and the sheets were air-dried at 37°C for 2 hours. Fixation of IgE (polyvinylpyrrolidone (PVP 40, Sigma, P0930) as blocking agent, human serum diluted 20 times) was revealed by luminescence (peroxidase-conjugated anti-human IgE (P 0295, Dako, Denmark, 100000 times diluted), SuperSignal West Dura Extended Duration substrate (34075, Pierce, Rockford, IL, USA) according to instructions of the manufacturer). Images were acquired using Luminescent Image Analyzer (LAS-3000, Fujifilm, Saint Quentin en Yvelines, France) in high resolution binning mode and 5min exposure time and were exported using multi Gauge software (ver 3.0). For competitive ELISA, digestion juices from pasta and pool of human sera (10 and 15 times diluted in PBS-T-G (0.5% gelatin (Sigma, G2500) in PBS buffer containing 0.1% Tween-20) respectively) were incubated for 3 hours at 37°C. Wheat proteins solution (purified α-gliadin or γ-gliadin or LMW glutenins ; albumin/globulin fraction ; 5µg/ml) were coated on 96 wells white plates Microfluor 2 (ThermoLabSystems, Franklin, MA, USA) for 2 hours at ambient temperature and blocked with PBS-T-G. Mixtures of digestion juices and human sera were then added for 15 hours at 37°C. Fixation of IgE was revealed by fluorescence using goat anti-human IgE alkaline phosphatase conjugate (ε-chain specific, Sigma A-3525) 500 times diluted and 4methylumbelliferyl phosphate (4-MUP) substrate (M-3168, Sigma) diluted 5 times. After incubation in the dark for 90min, fluorescence was measured at 440nm (excitation 360 nm) with FLX800 microplate reader and KC4 software (BioTek Instruments, Colmar, France). Data from in vitro digestion of cooked pasta and competitive ELISA were subjected to ANOVA and subsequent LSD test at the 5% significance using Statgraphics Plus 3.0 software (Manugistic Inc., Rochville, MD, USA). For ANOVA on in vitro digestion data, effects (pasta drying, pepsin time and pancreatin time) and all their interactions were first included then non-significant interactions (P>0.05) were excluded. Analysis in Principal Components of smoothed SE-HPLC profiles of digestion juices were performed in the Matlab environment using Matlab software (v7.0.4) (The MathWorks, Paris, France) and SAISIR (2008).
- 92 -
Chapitre 3. Pâtes alimentaires au blé dur
3. Results 3.1. Effect of the drying profile on the constituents of dried spaghetti
3.1.1. Starch
The gelatinization of starch was determined by DSC measurements. Table 3.1. shows the changes in DSC gelatinization endotherms from durum wheat semolina and dried spaghetti samples as a function of the drying profile. Whatever the drying profile used, dried pasta presented a lower gelatinization enthalpy (4.9J/g) compared to semolina (6.2 J/g) which is in accordance with previous studies (Guler et al., 2002; Vansteelandt et al., 1998; Yue et al., 1999; Zweifel et al., 2000). The reduced gelatinization enthalpy in dried pasta could be explained by the presence of partially gelatinized starch granules which require less energy to melt (Biliaderis, 1990). Moreover, in VHT pasta, the lower gelatinization enthalpy was accompanied by a higher onset (To) and peak (Tp) temperatures (56.4°C and 63.6°C respectively vs. 54.2°C and 61.8°C for semolina) and reduced temperature range (14.0°C vs. 16.1°C for semolina) as already observed (Guler et al., 2002; Vansteelandt et al., 1998; Yue et al., 1999). This may indicate that in VHT pasta, starch molecules had more mobility and underwent conformational reorganization conducting to an increase in crystallite stability and homogeneity as previously suggested by Yue et al. (1999). In comparison, other dried pasta underwent less marked changes.
Table 3.1. DSC measurements obtained from the first endothermic peak of DSC on dried spaghetti dried with one of the four drying profiles (LT, HT, VHT.LM, or VHT). Means (n=4) with the same superscript within a column are not significantly different (p>0.05). To, onset temperature; Tp, peak temperature; Tc; conclusion temperature, ∆Tr = Tc-To, temperature range, ∆H gelatinization enthalpy. Drying profiles Semolina
Dried pasta
To (°C) 54.2
LT
53.4
HT
53.5
VHT
56.4
VHT.LM
54.3
b b b a b
Tp (°C) 61.8 61.2 61.8 63.6 61.9
bc c bc a b
- 93 -
Tc (°C) 70.3 70.0 69.4 70.4 68.9
ab ab bc a c
∆Tr (°C) 16.1 16.5 15.9 14.0 14.6
b ab b c c
∆H (J/g db) 6.2 4.9 4.9 4.9 4.4
a b b b b
Chapitre 3. Pâtes alimentaires au blé dur
3.1.2. Proteins
Protein molecular weight distributions of durum wheat semolina and dried pasta were examined using SE-HPLC. Areas of SE-HPLC elution profiles from different protein extracts were used to measure protein extractability, which reflects protein aggregation. Figure 3.1 A shows the percentage of proteins soluble in SDS, in DTE followed by sonication and of the unextractable protein fraction. Durum wheat semolina was characterised by a high fraction of SDS-soluble proteins (81%), a low fraction of DTE-soluble proteins (19 %), and no traces of unextractable proteins. All dried pasta presented a higher protein aggregation as shown by the decrease in SDS-soluble proteins in favour of DTE-soluble proteins. This may indicate the formation of additional disulphide bonds. This phenomenon was enhanced with increasing drying temperatures, and particularly when the VHT was applied at a high pasta moisture content. Although all dried pasta presented a significant difference in protein solubility, 2 main groups of pasta could be distinguished: a group formed by LT and HT dried pasta (70 and 64% of proteins soluble in SDS; 28 and 35% soluble in DTE respectively) and another one formed by VHT.LM and VHT dried pasta (28 and 21% of proteins soluble in SDS; 64 and 68% soluble in DTE respectively). Moreover, the application of VHT.LM and VHT dryings resulted in the formation of other covalent bonds (e.g. isopeptide bonds), as shown by the presence of unextractable proteins in high proportions (10 and 12% respectively).
- 94 -
Chapitre 3. Pâtes alimentaires au blé dur
Figure 3.1. Peak areas of SE-HPLC elution profiles of SDS-soluble, DTE-soluble and unextractable protein fractions in semolina and dried pasta (A) and in cooked pasta (B). Means (n=3) with the same superscript within a graph (A or B) are not significantly different (P>0.05)
- 95 -
Chapitre 3. Pâtes alimentaires au blé dur
3.2. Effect of the drying profile on cooked spaghetti
3.2.1. Characterisation of the main constituents of cooked spaghetti
Starch After cooking at OCT+1, no gelatinization endotherm was observed on pasta whatever the drying profile used. In all pasta, starch was completely gelatinized as already observed (Colonna et al., 1990; Fardet et al., 1999).
Proteins The analysis of SDS-soluble, DTE-soluble and unextractable proteins in cooked pasta as a function of the drying profile is presented in Figure 3.1 B. Pasta cooking generated a decrease in SDS-soluble proteins in favour of DTE-soluble proteins and in a lesser extent of unextractable proteins. This phenomenon was especially pronounced for LT and HT cooked pasta in which cooking generated a decrease in SDS soluble fraction from 71% and 63% to 16% and 17%, respectively. In the same way, DTE-soluble fraction increased from 29% and 35 % to 72% for both cooked pasta. As a result, the cooking step levelled down the variation in protein solubility which were observed after drying except a small but significant lower SDS protein solubility in VHT and VHT.LM cooked pasta (about 14%). This lower proportion of SDS-soluble fraction is counterbalanced by a higher proportion of DTE-soluble and unextractable fractions. A more detailed analysis of SE-HPLC elution profiles from cooked pasta was then conducted. Elution profiles of SDS-soluble (Figure 3.2 A) and DTE-soluble fractions (Figure 3.2 B) were divided into five major fractions: S1 to S5 and P1 to P5 respectively. Elution profiles of SDSsoluble and DTE-soluble proteins in LT and HT cooked pasta were similar. Changes appeared to occur at higher drying temperatures. Compared to LT pasta, SDS-soluble proteins in VHT.LM pasta presented a significant higher proportion of fraction S1, and a lower proportion of fractions S3 S4 and S5. The same trend was observed in VHT pasta but in a lesser extent. Moreover, the DTE-soluble extract of VHT pasta presented a higher proportion of protein aggregates with a molecular weight between 62 kDa and 780 kDa (fractions P2 and P3) whereas in VHT.LM, the DTE-soluble extract was characterised by a high proportion (14%) of very large protein aggregates (> 116 kDa) (Figure 3.2 B).
- 96 -
Chapitre 3. Pâtes alimentaires au blé dur
Figure 3.2. SE-HPLC elution profiles (means=3) of SDS-soluble (A) and DTE-soluble (B) protein fractions in cooked pasta dried with one of the four drying profiles.
- 97 -
Chapitre 3. Pâtes alimentaires au blé dur
3.2.2. Pasta microstructure
Typical CLSM images of cooked pasta samples are shown Figure 3.3. Considering the staining procedure, the protein network appeared in white, whereas non fluorescent (dark) area could be mainly related to starch granules (Fardet et al., 1998a; Zweifel et al., 2003). Whatever the drying profile applied to fresh pasta, a continuous protein phase was visible. A method for objective description of protein network at microscopic scale was developed, based on digital image analysis, by evaluating the grey level granulometry. Erosion/dilation curves were built by successive application of opening and closing morphological operators (Figure 3.3 D). The left side (dilation) gives information on dark area, whereas the right side (erosion) reflects the protein film thickness, considering that one erosion steps makes the film loose 2 pixels, i.e. 0.8 µm (Devaux et al., 1997; Rouille et al., 2005). In order to provide an overall comparison of the images, the 108 erosion/dilation curves were processed by principal component analysis (PCA) (Figure 3.4 A). The first two principal components (PC) accounted for 97 % of the total variability. Whatever the drying profile, the external regions of pasta (coded “e”) were clearly distinguished from the central (coded “c”) and intermediate regions (coded “i”) by a lower PC1 score (Figure 3.4 A). In order to identify which parts of the erosion/dilation curve were contributing to the location effect, the first PC loading was examined (Figure 3.4 B). External regions of pasta were characterized by a lower amount of small dark area (< 6 µm) and a higher amount of larger dark area (> 6 µm) that could be related to a higher starch swelling. They were also characterised by a higher amount of the thinnest protein films (thickness lower than 4 µm). The protein distribution was therefore dependent on the location within pasta as already shown (Fardet et al., 1998b; Heneen et al., 2003; Zweifel et al., 2003). On the contrary to Fardet et al. (1998b) and Zweifel et al. (2003), but similarly to Heneen et al. (2003), similar protein distributions were observed in the intermediate and central regions of pasta. This difference could be explained by sample preparation for microscopic analyses that was similar to Heneen et al. (2003) but slightly different from Fardet et al. (1998b) and Zweifel et al. (2003).
- 98 -
Chapitre 3. Pâtes alimentaires au blé dur
Figure 3.3. Typical images (acquired by CLSM) of the LT cooked pasta in the central (A), intermediate (B), and external (C) zones and granulometric curves (D) corresponding to the three images
- 99 -
Chapitre 3. Pâtes alimentaires au blé dur
Figure 3.4. Principal component analysis of the granulometric curves (A) and first loading corresponding to principal component 1 (B).
- 100 -
Chapitre 3. Pâtes alimentaires au blé dur
For each location (central, intermediate and external zones), an effect of the drying profile could be observed. Whatever the location, the volume calculated from the CLSM images were lower in pasta dried at higher temperatures (HT, VHT and VHT.LM), implying a lower amount of proteins in these images (data not shown). In external regions, granulometric curves calculated from images of HT, VHT and VHT.LM drying profiles were significantly distinguished from LT drying profile (without any distinction between HT, VHT.LM and VHT dryings) (Figure 3.4 A). Similar trends were also observed for images acquired in the centre or intermediate zones of pasta. Pasta dried at high and very high temperatures (HT, VHT and VHT.LM) were characterized by a higher amount of the thinnest films (0.05) were excluded 2 Within the same row, the values with the same superscript are not significantly different (P> 0.05). For each analysed effect, mean value for all tested conditions for the other effects is indicated.
- 104 -
c
Chapitre 3. Pâtes alimentaires au blé dur
3.3.2. Effect of pasta drying on composition of digestion juices
Digestion juices were analyzed by SE-HPLC during in vitro digestion. SE-HPLC profiles of digestion juices at 220nm are illustrated in Figure 3.5 A for LT-pasta. As expected, digestion shifted profiles toward lower molecular weight and increased amount of material in digestion juices. SE-HPLC profiles of digestion juices at 220nm were then analysed by PCA (Figure 3.5 B and C) and ANOVA were performed on the principal components. On the first plan of the PCA (Figure 3.5 C), SE-HPLC profiles of digestion juices gathered as a function of the digestion step. On the first dimension (80% of variance), large differences between profiles from each pancreatic time (30 and 180 min) and the group of profiles from the end of buccal and gastric phases were found. The second dimension (13% of variance) distinguished profiles of the end of the buccal phase (-0.224) from profiles of the end of the gastric phase (0.205) and from the two sets of profiles from the pancreatic phase (0.043 and –0.024 for 30 minutes and 180min of pancreatic digestion respectively). Only profiles of pancreatic digests were significantly influenced by the drying profile resulting in lower and higher first component respectively for VHT and VHT.LM profiles (Figure 3.5 C). Taking into account that higher first component value appeared linked to lower digestion, this result agrees with lower DH values previously found for VHT.LM pasta (29% and 55% at 30 and 180 min, respectively) compared to 33% and 65% at 30 and 180 min, respectively for VHT. No significant effect of the drying profile was found on the second component. Both quantitative and qualitative differences in the composition of polypeptides in digestion juices could result in such differences in SE-HPLC profiles. Pancreatic digestion juices of VHT and VHT.LM pasta were then further analysed by reversedphase HPLC but no qualitative differences could be found (results not shown), only quantitative differences appeared to be responsible for differences found on SE-HPLC profiles of digestion juices.
- 105 -
Chapitre 3. Pâtes alimentaires au blé dur
Figure 3.5. SE-HPLC profiles at 220 nm of digestion juices of LT-pasta in the course of in vitro digestion (A) and (B) loadings and (C) first plan of the Principal Components Analysis (PCA) of SE-HPLC profiles of pasta at the end of buccal phase, gastric phase, after bucco-gastric phase + 30 min of pancreatic phase and after bucco-gastric-phase +3h of pancreatic phase.
- 106 -
Chapitre 3. Pâtes alimentaires au blé dur
3.3.3. Effect of pasta drying on residual allergenicity of digestion extracts and juices
Wheat proteins from pasta were first investigated for their IgE binding during digestion by dot blotting (data not shown). As expected, IgE binding was detected with the serum from a patient allergic to based wheat food, whereas no signal was detected for the patient allergic to Graminaceae excepted a faint signal for the albumin/globulin fraction control. Whatever the drying profile, IgE binding to extracts decreased as digestion proceeded to be hardly detected at the end of the pancreatic phase in tested conditions. However, at all tested steps of digestion, IgE binding was still detected in digestion juices. Juices progressively concentrated solubilized proteic fragments, some of which still kept IgE binding capacities that could be at risk for allergic patients whenever there were not fully processed by peptidases from the brush border. These results show that IgE-reactive peptides that resist to the digestion process are formed whatever the drying profile in agreement with previously reported data (De Zorzi et al., 2007). To compare IgE reactivity of proteic fragments in digestion juices from different pasta and to get an insight of the proteic origin of these fragments, competitive ELISA was then performed with digestion juices and a pool of sera of 4 patients suffering from food allergy to wheat against wheat proteins fractions. Higher concentration of IgE-reactive peptides in digestion juice will lead to higher inhibition of wheat protein recognition provided that IgE interact with preserved epitopes of tested protein. Linked to IgE reactivity of the pool against wheat protein, competitive ELISA was performed for the prolamin fraction against α and γ-gliadins and low molecular weight glutenins and against the albumin/globulin fraction. As shown in Table 3.4, inhibition varied depending on wheat fraction. Direct comparison of inhibition percentage must be restricted to proteins with similar abundance in semolina and IgE reactivity ; α-gliadin and albumins/globulins fraction meet these conditions. Inhibition by digestion juice of LT pasta was thus found higher for α-gliadin than for soluble albumins/globulins at the end of gastric phase (66 and 28 %) but comparable at the end of the intestinal phase (16 and 14 %). Reduction of IgE reactivity with digestion also depended on wheat fraction: for LT pasta for example, reduction of inhibition by a factor 4 for α-gliadin, 2 for albumins/globulins fraction and low molecular weight glutenins but rather no reduction was found for γ-gliadin. This last result was found whatever the drying profile. Compared to LT pasta, drying process led to reduced allergenicity of digestion juices from the end of the gastric phase in case of VHT profiles (lower inhibition of α-gliadin recognition for VHT and VHT.LM pasta and of low molecular weight glutenins recognition for VHT.LM pasta). For digestion juices from the end of the intestinal phase, recognition of gliadins was not modified by drying process but allergenicity was increased with higher inhibition of low - 107 -
Chapitre 3. Pâtes alimentaires au blé dur
molecular weight glutenins recognition for HT and VHT.LM pasta and of albumins/globulins fraction for VHT.LM pasta. VHT.LM pasta exhibited the lowest reduction of allergenicity upon digestion for α-gliadin (by 2) and albumins/globulins fraction (by 1). To conclude on these results, digestion juices from pasta were found to inhibit recognition of each tested wheat fraction by IgE from allergic patients showing that these juices gather proteic fragment with allergenic potentialities originated from different wheat fractions. Impacts of drying profile on allergenic properties of digestion juices depended on digestion step and on wheat protein which recognition is involved that is to say that drying process can lead to various results (increase, decrease, no change) depending on sensibilization profile of allergic patients but no tested drying profile was found to abolish allergenic properties of wheat proteins.
Table 3.4. Results of competitive ELISA with digestion juices from pasta and a pool of sera from allergic patients to wheat. Percentage of inhibition obtained with digestion juices at the end of the gastric phase (5 minutes of hydrolysis by α-amylase and 3 hours by pepsin) and at the end of the intestinal phase (5 minutes of hydrolysis by α-amylase, 3 hours by pepsin and 3 hours by pancreatin) are presented. Means (n=2) with the same superscript by protein are not significantly different (P > 0.05). Fluorescence intensity without inhibitor α-gliadin
γ-gliadin
Low MW glutenins Albumin/ globulin fraction
Digestion juice from the end of
LT
HT
VHT
VHT.LM
gastric phase
66 a
64 a
44 b
40 b
intestinal phase
16 cd
14 cd
10 d
19 c
gastric phase
88 a
79 bc
80 abc
81 ab
intestinal phase
78 bc
75 bc
72 c
77 bc
gastric phase
93 a
91 ab
90 ab
87 b
intestinal phase
51 d
66 c
46 d
63 c
gastric phase
28 a
26 ab
24 ab
24 ab
intestinal phase
14 c
16 c
14 c
21 b
31753
9989
12118
31732
- 108 -
Chapitre 3. Pâtes alimentaires au blé dur
4. Discussion The first objective of this study was to investigate the effects of pasta drying on the structure of dried and cooked pasta, characterised by a multi-scale approach. Pasta dried at low temperature (LT) showed small but significant differences in starch and protein structures, when compared to semolina. Starch presented a reduced gelatinization enthalpy, which may indicate the presence of partially gelatinised starch. LT dried pasta was also characterised by a higher protein aggregation probably through disulphide bonds (protein solubility in SDS was decreased by 10% in favour of DTE solubility). An increase in the drying temperature up to 70°C had a moderate impact on starch and protein structures. However, the application of a very high temperature (VHT and VHT.LM), especially at high pasta moisture content (VHT), induced significant reorganization of starch and protein fractions. The results suggested that starch from VHT pasta presented a more stable and homogeneous crystalline structure. VHT.LM and VHT drying induced a high protein aggregation, as already shown on pasta (Aktan et al., 1992; Lamacchia et al., 2007; Zweifel et al., 2003). This protein aggregation probably occurred through the formation of additional disulphide bonds and other covalent bonds (i.e. isopeptide bonds).
Further structural changes occurred during the cooking step. On the basis of DSC measurements, starch was totally gelatinized in all cooked pasta. This was not surprising since they were cooked until the disappearance of the white core corresponding to ungelatinized starch. Cooking led to a marked protein aggregation in LT and HT pasta, probably through disulphide bonds but also through other covalent bonds. In contrast, proteins from VHT and VHT.LM pasta underwent less marked changes during cooking probably because most of the proteins were already aggregated during the previous drying step. Cooking levelled down differences in protein structure created by the different drying profiles and led to an apparent similarity between cooked pasta. However, the analysis of protein size distribution revealed that VHT.LM cooked pasta had a higher proportion of larger protein polymers. This small difference seen at a molecular level may induce important changes at a macroscopic one.
Indeed, an increase in drying temperature corresponded to an increase in pasta hardness, cohesiveness, resilience and springiness (TPA test) and in breaking energy (tensile test). In particular, VHT.LM drying induced an increase by 15% in pasta hardness and by 40% in pasta breaking energy compared to LT drying. It suggests that more severe drying conditions may
- 109 -
Chapitre 3. Pâtes alimentaires au blé dur
promote the formation of a strong protein network responsible for the higher resistance of pasta to compression and tensile forces.
The microstructure of all cooked pasta was characterised by large structural differences between the external and the central regions. This result can be a consequence of the gradient in moisture distribution created during cooking (Baiano et al., 2006; Horigane et al., 2006). The core and intermediate regions of cooked pasta presented a dense protein network surrounding voids assimilated to starch granules with varying sizes. The external zone was characterised by a looser protein network surrounding larger voids assimilated to highly swollen starch granules. This is in accordance with previous studies on pasta (Cunin et al., 1995; Heneen et al., 2003; Zweifel et al., 2003). The increase in drying temperature did not create much difference in pasta microstructure.
The second objective of our study was to determine whether structural changes, induced by pasta dryings, could modify the digestibility of protein and starch fractions and to know whether structural changes in the protein fraction induced by pasta drying and/or the digestion process could abolish the presence of IgE-reactive peptides. The bucco-gastric digestion of cooked pasta led to 22% of starch hydrolysis and a limited protein hydrolysis (8%), whatever the drying profile. The effect of the drying profile on protein and starch digestibility was more pronounced at the duodenal phase. In particular, in VHT.LM cooked pasta protein and starch hydrolysis was reduced by 10% and 3% respectively after three hours of intestinal phase, in comparison with LT and HT pasta. VHT.LM cooked pasta was also characterised by a high proportions of protein aggregates in protein extracts and of larger peptides in digestion juices. Moreover, none of the drying profile tested was found to abolish allergenic properties, digestion juices from VHT.LM pasta were even found to contain more IgEreactive peptides from albumins/globulins fraction. These results are in agreement with two previous studies on pasta where drastic thermal treatment were shown to (1) decrease the in vitro digestibility of starch (Casiraghi et al., 1992) and (2) reduce the in vitro digestibility of wheat proteins without affecting their potential allergenicity (De Zorzi et al., 2007).
The application of a very high temperature, especially at low pasta moisture content (VHT.LM) has contributed to the formation of a different structure, certainly responsible for the lower protein and starch hydrolysis. This difference was not observed at a microscopic scale but at a molecular one. No major changes were observed on the spatial distribution of proteins. - 110 -
Chapitre 3. Pâtes alimentaires au blé dur
However, molecular rearrangements of proteins have led to the formation of large protein aggregates which have probably contributed to the moderate decrease in the in vitro protein and starch hydrolysis and increase in the residual allergenicity. However, it would be interesting to confirm these results in vivo.
Acknowledgements This work was carried out with the financial support of the « ANR- Agence Nationale de la Recherche - The French National Research Agency » under the « Programme National de Recherche en Alimentation et nutrition humaine », project « ANR-05-PNRA-019, PASTALEG». We are indebted to B. Chabi from the Montpellier RIO Imaging Platform (RIO Imaging facility, Montpellier, France) for imaging experiments and B. Vernus from DCC laboratory (UMR DCC, Montpellier, France) for microtome use. The authors are very grateful to B. Chauvet (UR BIA, Nantes, France) for in vitro digestion experiments and to J. Bonicel (UMR IATE, Montpellier, France), G. Dehayes and M.A. Legoux (UR BIA, Nantes, France) for their technical assistance in SE-HPLC analyses and dot-blotting experiments. The authors wish to thank B. Bouchet, F. Guillon (UR BIA, Nantes, France), J. Abecassis and B. Cuq (UMR IATE, Montpellier, France) for helpful discussions about microscopic analyses and pasta processing.
- 111 -
Chapitre 3. Pâtes alimentaires au blé dur
Chapitre 3: CONCLUSIONS Cette étude a permis de montrer que l’application de différents traitements de séchage sur des pâtes au blé dur engendrait des modifications structurales affectant les protéines et l’amidon. Dans les pâtes sèches, cela se traduit par une agrégation importante des protéines dans les pâtes séchées à très haute température (VHT et VHT.LM), en particulier lorsque la très haute température est appliquée en début de cycle (VHT). Cette agrégation implique la formation de ponts disulfures additionnels et d’autres types de liaisons covalentes fortes. Concernant l’amidon, l’application du traitement VHT engendre une réorganisation structurale de l’amidon, augmentant sa stabilité et son homogénéité. Dans les pâtes cuites, l’analyse des propriétés rhéologiques montre une augmentation de la fermeté, de la cohésion, de l’élasticité et de la résistance à la rupture avec l’augmentation des températures de séchage, en particulier lors de l’application d’un séchage très haute température en fin de cycle (VHT.LM). Cela traduit un renforcement de la structure, probablement lié au réseau protéique. Au niveau microscopique, peu de différences de distribution du réseau protéique sont observées quel que soit le séchage appliqué. Au niveau supramoléculaire, les différences structurales engendrées par les différents séchages sont nivelées par la cuisson. La cuisson conduit à la gélatinisation totale de l’amidon et à une agrégation massive des protéines due en majeure partie à la formation de ponts disulfures mais également d’autres liaisons covalentes. Cependant une analyse plus détaillée de la distribution en taille des protéines permet de mettre en évidence la présence de gros agrégats protéiques non réduits dans les pâtes séchées à très haute température, notamment lorsque la haute température est appliquée en fin de cycle (VHT.LM). Au niveau nutritionnel, l’application d’un séchage VHT.LM engendre une diminution (-10%) de l’hydrolyse des protéines et tend à diminuer (-3%) la digestibilité de l’amidon en fin de phase intestinale. La diminution de digestibilité protéique et glucidique ne peuvent être expliquées ni par une différence d’épaisseur du réseau protéique au niveau microscopique ni par la présence d’amidon non gélatinisé. Elle pourrait être dûe à la présence d’une proprotion plus grande de gros aggrégats protéiques au niveau supramoléculaire, témoignant d’un réseau protéique particulier. Par ailleurs, la présence de fragments allergènes est toujours détectable quel que soit le séchage subi ou la phase de digestion considérée. Comparé aux autres traitements, le séchage VHT.LM conduit à une plus forte réactivité des jus de digestion pancréatiques (en fin de phase intestinale) des pâtes vis-à-vis des IgE de patients allergiques au blé reconnaissant les gluténines de haut-poids moléculaire et la fraction albumines-globulines.
- 112 -
Chapitre 4 Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
- 113 -
- 114 -
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
Chapitre 4 : INTRODUCTION
Le chapitre 3 a pu mettre en évidence que des changements structuraux mineurs, engendrés par divers traitements de séchage, pouvaient modifier les propriétés nutritionnelles des pâtes au blé dur. Des changements structuraux pourraient être aussi induits par une modification de la formulation et notamment par l’introduction de farine de légumineuse. Le développement de produits mixtes « blé dur/légumineuse » de haute qualité nutritionnelle, et notamment de pâtes enrichies en farine de légumineuse, connaît depuis quelques années un intérêt croissant. Les taux d’enrichissement varient généralement de 5 à 30%. Une seule étude a été conduite sur des pâtes enrichies avec 50% de farine de pois chiche. Cependant, dans la plupart de ces études, les pâtes sont fabriquées à l’échelle laboratoire. Elles ne permettent donc pas de prédire la bonne processabilité à l’échelle pilote ou industrielle. Ce chapitre est consacré à l’étude de la processabilité à l’échelle pilote de pâtes enrichies avec une quantité élevée (35%) de farine de fève ou de pois cassé, ainsi qu’à la caractérisation de leurs propriétés culinaires.
- 115 -
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
- 116 -
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
Publication 4 : Fortification of pasta with split pea and faba bean flours : pasta processing and quality evaluation Maud Petitot, Valérie Micard
Food Research International, sous presse.
ABSTRACT Nutritionally enhanced spaghetti was produced by adding high amounts (35% db) of legume flour (split pea or faba bean) to durum wheat semolina. The production of fortified pasta required an adaptation of the pasta making process (higher hydration level and mixing speed) to limit the agglomeration of particles during mixing. Moreover, legume flour addition induced a decrease in some pasta quality attributes (e.g. higher cooking loss, lower breaking energy). This could be attributed to the introduction of non gluten proteins and insoluble fibres which weakened the overall structure of pasta. A modification of the sensorial properties including higher hardness and higher fracturability were also observed. Some of the quality attributes of fortified pasta were improved (e.g. lower cooking loss) by applying high and very high temperatures during the drying cycle, reflecting the protein network strengthening. However, these treatments resulted in excessively firm and rubbery pasta according to the panellists.
Key-words: Durum wheat pasta; Legumes; Cooking quality; Sensory evaluation
- 117 -
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
1. Introduction Pasta products, largely consumed all over the world are traditionally manufactured from durum wheat semolina, known to be the best raw material suitable for pasta production (Feillet et al., 1996). Pasta is a source of carbohydrates (74-77%, db) whose interest is increasing due to its nutritional properties, in particular its low glycaemic index (GI) (Monge et al., 1990). Pasta contains also 11-15% (db) of proteins but as most cereals products, pasta is deficient in lysine (the most deficient amino-acid) and threonin (the second limiting amino-acid) (Abdel-Aal et al., 2002; Kies et al., 1970). That’s why non traditional raw materials can be used in order to increase the nutritional quality of pasta (Brennan et al., 2003; Chillo, Laverse, Falcone & del Nobile, 2008a). Among these non traditional raw materials, legumes represent an interesting source of proteins, fibres, vitamins and minerals. Legume proteins are relatively low in sulphur-containing amino-acids, methionine, cysteine and tryptophan but high in lysine. As a consequence, legumes and cereals are nutritionally complementary (Duranti, 2006): legume flours can be used to partially substitute durum wheat semolina in pasta. However, depending on the substitution level, the pasta making process could be more or less difficult. Some authors reported that pasta production above 30% fortification with chickpea flour was not possible because of the aggregation of particles during mixing which renders the feeding of the extrusion screw difficult (Wood, 2009). Moreover, high level of substitution could decrease the quality of pasta products. Up to a 10% substitution level, pasta fortified with lupin flour (Rayas-Duarte et al., 1996; Torres et al., 2007a) or chickpea flour (Wood, 2009; Zhao et al., 2005) are generally well accepted but at higher level of substitution, negative changes in the cooking quality (higher cooking loss and stickiness) and sensorial attributes (poor acceptability) appear (Bahnassey et al., 1986a; RayasDuarte et al., 1996; Torres et al., 2007a; Zhao et al., 2005). Indeed, the introduction of legume flour dilutes the gluten and weakens the protein network which is responsible for example for leaching of more solids in the cooking water (Rayas-Duarte et al., 1996; Torres et al., 2007a). Application of high drying temperatures in pasta processing was shown to increase the quality of durum wheat pasta (lower cooking loss, lower stickiness, higher firmness) (Baiano et al., 2006; Zweifel et al., 2003). High drying temperatures could thus be a way to improve the culinary and sensorial properties of fortified pasta. The objectives of this work were to determine the impact of a high level of legume addition on the pasta manufacturing process and on the culinary and sensorial attributes of these new pasta products. The effect of high drying temperatures on pasta quality was also studied. - 118 -
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
Split pea and faba bean flours have been selected for this study as they received less attention than other legume flours such as lupin and chickpea.
2. Materials and Methods 2.1. Raw materials
Durum wheat semolina was supplied by Panzani (Marseille, France). Dehulled split pea and faba bean flours were obtained from Terrena (Moisdon la Rivière, France).
2.2. Chemical composition of semolina and flours
Total protein content was determined using Kjeldahl procedure with a nitrogen-to-protein conversion factor of 5.7. Total starch content was determined with an enzymatic assay kit (Megazyme, Co. Wicklow, Ireland). Ash analysis used an incineration at 900°C for 2h according to a French norm (NF 03-720). Lipid content was determined according to the method of Bligh & Dyer (1959). Fibres were determined by ISHA (Lonjumeau, France) according to the AOAC 985.29 method. Vitamin (B1, B2, B3, B6, and B9) and mineral (iron, magnesium, phosphorus) compositions were determined by INZO (Chateau-Thierry, France). Analyses were conducted in triplicate except for fibres, vitamins and minerals (duplicate).
2.3. Particle size distribution of semolina and flours
Particle size distribution of durum wheat semolina, split pea and faba bean flours were analysed in triplicate by a laser diffraction particle size analyzer Coulter LS 230 (Beckman Coulter Inc, Fullerton, USA).
2.4. Ability of legume flours to be processed into pasta 2.4.1 Granulation properties of durum wheat semolina and legume flours: minimum water content for dough formation
The minimum water content for dough formation with durum wheat semolina and blends of 65% durum wheat semolina and 35% legume flour were evaluated during mixing under - 119 -
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
continuous water addition as described by Landillon, Cassan, Morel & Cuq (2008) using a Farinograph small volume mixing bowl (Brabender, OGH, Duisburg, Germany) connected to a Plasticorder PL200 drive system (Brabender, OGH, Duisburg, Germany) equipped with a torquemeter. A 50 g sample was hydrated at a constant rate (1ml.min-1) and mixed at a constant speed (63 rpm). Mixing torque vs. water content was plotted. The minimum water content for dough formation (% db) was recorded. It corresponds to the critical water content at which a rapid increase in torque is measured. Analyses were done in duplicate.
2.4.2. Pasta manufacturing
Durum wheat pasta (control): spaghetti was processed as described by Petitot et al. (2009c) with a continuous pilot-scale pasta extruder (Bassano, Lyon, France). Durum wheat semolina (5 kg) was hydrated with tap water to obtain a moisture content of 47% (db) (or 32% wb) and then mixed for 15 min at 60 rpm. The product was then extruded at 31 rpm and 40°C. Extruded spaghetti was then dried at low temperature 55°C (LT) in a pilot-scale drier (AFREM, Lyon, France) in order to reach 12% of moisture. The relative humidity and temperature inside the dryer and the moisture content of pasta (evaluated according to AACC method (44-15)) were monitored during drying. The diameter of dry spaghetti was 1.56 mm ± 0.02. Durum wheat pasta dried at LT was considered as a control. Pasta fortified with split pea or faba bean flour: durum wheat semolina was substituted by 35% (db) of legume flour (split pea or faba bean). Based on the study of the granulation properties of durum wheat semolina and legume flours blends and the operator experience, a hydration level of 44% (db) (or 31% wb) was considered optimal and was chosen for the production of fortified pasta. 35% legume fortified pasta was processed as described for durum wheat pasta except that three minor changes were applied 1) semolina and legume flour were first mixed together for 5 min before hydration, 2) the optimal hydration level was reduced to 44g/100g db, and 3) the mixing speed was increased to 120 rpm for 15 min in order to limit the formation of particle aggregates. The product was then extruded as for durum wheat pasta. Extruded spaghetti was dried in a pilot-scale drier (AFREM, Lyon, France) in order to reach 12% of moisture. Three drying profiles were applied: low temperature 55°C (LT), high temperature 70°C (HT), very high temperature 90°C applied at the end of the drying cycle, when the moisture content of pasta is low (about 12%) (VHT.LM). The diameter of all dry spaghetti was 1.56 mm ± 0.02.
- 120 -
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
2.4.3. Particle size distribution of the dough after mixing and pressure during extrusion Agglomeration properties of durum wheat semolina and legume flours during mixing were evaluated by sieving 100 g of the dough through 10 mm, 6.3 mm, 4 mm, 3.15 mm, 2 mm and 1 mm mesh screens (Landillon et al., 2008). The percentage of three different fractions: small particles (d < 1mm), agglomerates (1 < d < 4 mm) and dough pieces (d > 4mm) was calculated. Extrusion pressure at the head of the extrusion press (at the die extruder entrance) was monitored by a pressure indicator (Gentran inc., Fremont, USA) and was recorded when it was stable. Results were expressed as the mean of 6 replications of pasta production ± SD.
2.5. Spaghetti cooking quality 2.5.1. Cooking time Dried spaghetti was cooked in Evian water (2L / 100g) containing 0.7% (w/v) of sodium chloride. Optimal cooking time (OCT) was indicated when the white core of the pasta disappeared when squeezed between two glass plates (approved method 66-50 AACC 2000). All analyses on cooked pasta were made on pasta cooked at OCT+1min. 2.5.2. Water uptake in cooked pasta Water uptake was evaluated on a 50 g dry pasta sample and calculated using the equation:
Water uptake (% db) =
weight of cooked pasta (OCT + 1 min) − weight of dry pasta × 100 weight of dry pasta
2.5.3. Cooking loss The dry matter (Dm) content of a dry pasta sample was first determined. Then the pasta sample was cooked at OCT+1min and its dry matter content was determined by a two-stage drying procedure. First, a pre-weighed cooked pasta sample was dried in an oven at 50°C for 2 days. The sample was then ground and an aliquot was weighed and dried in oven at 130°C for 2 hours. The cooking loss was calculated: Cooking loss (%, db) =
Dm of cooked pasta (g ) − Dm of dry pasta (g ) × 100 Dm of dry pasta (g )
- 121 -
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
2.6. Colour of dry and cooked spaghetti Colour of dry and cooked spaghetti was determined by a Minolta Chromameter (Model CR-400, Minolta Co., Osaka, Japan) using the Hunter L*, a*, b*. L* values measure black to white (0-100); a* values measure redness when positive and greenness when negative; b* values measure yellowness when positive and blueness when negative. Each data colour represents the mean of 3 measurements on 3 different pasta samples.
2.7. Rheological properties of cooked spaghetti A TA-XTplus (Stable Micro Systems, Scarsdale, USA) texture profile analyser equipped with a windows version of Texture Expert software package (Stable Micro Systems, Scarsdale, USA) was used to evaluate rheological properties of cooked spaghetti. After cooking (at OCT+1min), pasta was allowed to equilibrate at ambient temperature for 20 min in a covered container at 25°C before texture analysis. The variables were recorded through five measurements on each pasta sample, cooked in two different occasions, totalling to 10 measurements for each pasta sample.
2.7.1. Texture profile analysis
The TA-XTplus was equipped with a 35 mm cylindrical probe (ref. P/35, Stable Micro Sytems). The probe compressed a single strand of cooked spaghetti at a constant rate of deformation (1 mm/s) to 70% of the initial spaghetti thickness. The probe was retracted and held stationary 10 s before performing a second compression to 70% of the original spaghetti thickness. From texture profile analysis curve, textural parameters of hardness, cohesiveness, and resilience were obtained (Epstein et al., 2002). Spaghetti hardness was defined as the maximal peak force attained during the first compression. Cohesiveness was calculated as the ratio of the area under the second peak to the area under the first peak. Resilience was defined as the ratio of the area under the second half of the first peak to the area under the first half of the same peak.
- 122 -
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
2.7.2. Tensile test
The TA-XTplus was equipped with tensile grips (ref. A/SPR, Stable Micro Sytems). The initial distance between the two tensile grips was 15 mm. The test was performed at a constant rate of deformation (3 mm/s). The stress-strain curve was plotted and the energy (J.m-3) stored in the sample until fracture, which corresponds to the area under the curve, was calculated.
2.8. Sensory evaluation Sensory evaluation was carried out on cooked spaghetti (OCT+1min) by a panel of 14 trained judges from the European Research Centre of Panzani (Crecerpal, Marseille, France) using a descriptive analysis technique. Pasta samples were presented to each judge separately. The order of sample presentation was randomized. The judges were asked to score different quality characteristics, including appearance and texture on a 10-points scale where 1 represented low intensity and 10 represented high intensity in a particular attribute. The following quality attributes were evaluated: surface smoothnes (surface property which indicates absence of cracks and roughness); homogeneity (difference in texture between the surface and the core of a spaghetti strand) (1: heterogeneous; 10: homogeneous). Hardness: the force required to achieve a given deformation or penetration (NF ISO 5492, 1992). (1: soft; 5: firm, 10: hard). Springiness: the degree to which the deformed material returns to its initial condition after the deforming force is removed (NF ISO 5492, 1992) (1: plastic; 5: elastic; 10: rubbery). Fracturability: the force with which the product crumbles. It is the result of a high hardness and a low degree of cohesiveness (NF ISO 5492, 1992) (1: crumbly; 5: crunchy; 10: brittle).
2.9. Statistical analyses Data from dough particle size distribution after mixing, extrusion pressure, cooking quality, colour, TPA test and sensory evaluation were subjected to analysis of variance (ANOVA) followed by the Fisher’s least significant difference (LSD) test to compare means at the 5% significance level by using Microsoft Xlstat software 2008 (Addinsoft, Paris, France). Data from the tensile test were subjected to Kruskal-Wallis test followed by the Dunn multiple comparison test to compare means at the 5% significance level by using Microsoft Xlstat software 2008 (Addinsoft, Paris, France).
- 123 -
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
3. Results and discussion 3.1. Chemical composition of semolina and flours The two main constituents of durum wheat semolina are starch (~78%) and proteins (~13%) (Table 4.1). Durum wheat semolina contains also small amounts of fibres, lipids, vitamins and minerals. The introduction of legume flour in pasta increases the protein, insoluble fibre, vitamin (B1, B5, B6, B9) and mineral (Fe, Mg, P) contents. Moreover, legume proteins are rich in lysine and threonin (Alonso et al., 2000; Carbonaro et al., 1997), two essential amino acids that are limiting in durum wheat (Abdel-Aal et al., 2002; Kies et al., 1970). The introduction of legume flour could thus increase the nutritional quality of pasta, especially if high levels of legume flour are added. However, this could impact on processing ease and quality attributes of pasta, as it will be presented.
3.2. Ability of legume flours to be processed into pasta
The study of granulation properties of semolina and legume flours revealed that the minimum water content for dough formation was lower when semolina was substituted with 35% of split pea or faba bean flour (Table 4.2). From this result and from operator experience, it appeared that a reduction in the hydration level from 47% to 44% (db) was suitable for the production of 35% legume fortified pasta. Moreover, the mixing speed was increased (from 60 to 120 rpm) in order to limit the formation of aggregates and dough lumps. However, even by controlling the hydration level and mixing speed, it appeared that the introduction of 35% of legume flour was the highest proportion that could be reached because of particle aggregation. Indeed, the characterization of particle size distribution after the mixing step revealed that a substitution of durum wheat semolina with 35% of legume flour induced the formation of heterogeneous dough with larger lumps, especially for the dough containing 35% of faba bean flour (Table 4.2). Durum wheat dough was characterized by a high amount of small particles (< 1 mm) (68%), and almost no large lumps (> 4 mm). In contrast, the dough made from 35% of legume flour presented a lower amount of fine particles ( 4 mm) (8-14%).
- 124 -
Chapitre 4. Pâtes alimentaires enrichies en légumineuse : processabilité et qualité culinaire
Table 4.1. Chemical composition of durum wheat semolina, split pea flour, faba bean flour and blends of 35% legume flour and 65%
durum wheat semolina. Protein, starch, fiber, lipid and ash contents are expressed in g/100g (db); vitamin and mineral content is expressed in mg/100g (db) Fibres Proteins
Starch
Vitamins
Minerals
Ashes
Lipids Total
Soluble
B1
B2
B3
B5
B6
B9
Fe
Mg
P
DW semolina
13.3 ± 0.2
77.6 ± 0.3
1.7 ± 0.04
2.4
0.7
1.1 ± 0.02 0.28 50 nm) and mezopores (2-50 nm) was obtained from the curve of the cumulative intrusion volume vs. pore size.
2.4. Water absorption kinetic during cooking Water absorption tests were performed according to the method of del Nobile, Baiano, Conte Mocci (2005). Culture tubes containing 9 ml of distilled water were equilibrated at 100°C ± 1°C in a thermostated bath. 40 ± 0.5 mm spaghetti samples were immersed in the tubes, one for each tube. At given time (every 30 s for the first 10 min, every minute up to 15 min), samples - 144 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
were removed from the tubes, rapidly blotted and accurately weighed. 3 determinations (3 tubes containing one pasta strand each) were made for each sample. The ratio between the increase in weight and the weight of dry spaghetti (∆W/W0) = W(t) – W0/W0 was plotted in function of time.
2.5. Microstructure
2.5.1. Characterisation of wheat and legume starch granules Extraction procedure: - Durum wheat starch: 100g of semolina was mixed with water (1:1) using a Kenwood mixer (KM221, Kenwood, France) until the formation of a sticky dough. Then, water was added (200 ml) and mix again for 1min in order to extract starch from the gluten matrix. The suspension was passed through a set of sieves (200, 160, 120, 75 and 50 µm) using a sieve shaker (Retsch AS 200 digit, GmbH & Co, Germany) with an electromagnetic drive connected to tap water. The filtrate thus obtained was centrifuged (3000g/20min/4°C) (Beckman Avanti centrifuge TM J-30 I, Fullerton, USA) and the yellow-brown upper layers containing the tailings starch (rich in protein) was carefully separated from the bottom layer containing the prime starch. The prime starch was purified by successive washing steps in water and centrifugations (3000g/20min/4°C). This process was repeated until no “sludge” layer could be detected. Starch was air-dried overnight and stored at room temperature. - Legume starch: 100g of split pea or faba bean flour was mixed with water (1:6) adjusted to pH 9 with 1M NaOH in order to increase protein solubility according to Carbonaro et al. (1997). The suspension was sieved and the filtrate thus obtained was centrifuged (3000g/20min/4°C). Prime starch was then purified as for durum wheat starch. Particle size distribution Durum wheat semolina, split pea and faba bean starch granules were analysed in triplicate by a laser diffraction particle size analyzer Coulter LS 230 (Beckman Coulter Inc, Fullerton, USA). Environmental scanning electron microscopy (ESEM) ESEM (Xl 30 ESEM, Philips, Netherlands) equipped with a gaseous secondary electron detector was used to observed native starch granules. Purified starch samples were mounted on stubs using double-sided adhesive tape. Samples were observed at ambient temperature with a 20kV acceleration voltage and at a pressure of 2 Torr.
- 145 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
2.5.2. Microstructure of cooked spaghetti
2.5.2.1. Sample preparation After cooking (OCT+1), pasta was rinsed with cold tap water, drained for 1 min, and equilibrated in a covered container at 25°C for 15 min before sectioning. Pasta was cut in small pieces and frozen quickly at -40°C in water plus a cryoprotector (Cellpath, Newtown, UK). Samples were cut at -20°C using a microtome (Microm HM 560, Walldorf, Germany). 8µm thick cross sections were used for bright field, florescence and polarised light microscopy whereas 15 µm thick cross sections were used for confocal laser scanning microscopy.
2.5.2.2. Bright field light microscopy Proteins were stained with 1g.L-1 Fast Green (Sigma Aldrich Co., USA) and starch granules with iodine by applying Lugol’s solution (Fluka, Buchs, Switzerland) diluted 1:8 (v/v) for 1 min. The proteins stained green while starch granules stained a blue-brownish color. Images were observed using a Leica DM6000M microscope equipped with a Leica DFC290 digital camera and Leica Qwin image software (version 3.0) (Leica Microsystems, Wetzlar, Germany). 2.5.2.3. Polarised light microscopy (PLM) Loss of birefringence was detected under polarised light. Unstained images were observed using an Olympus BX61 microscope (Olympus America Inc., Center Valley, USA) equipped with a Hamamastu ORCA-AG camera (Hamamatsu Photonics K.K, Hamamatsu, Japan) and the Cell^P software (Olympus America Inc., Center Valley, USA). 2.5.2.4. Fluorescence microscopy Fibres were stained with 0.1g.L-1 Calcofluor White M2R in 0.2 mol L-1 phosphate sodium buffer (pH 8.0) for 10 min, counterstained by 1g.L-1 Fast Green. Stains were purchased from Sigma-Aldrich Co. (USA). Images were observed using an epi-fluorescence Olympus BX61 microscope (Olympus America Inc., Center Valley, USA), equipped with a Hamamastu ORCAAG camera (Hamamatsu Photonics K.K, Hamamatsu, Japan) and the Cell^P software (Olympus America Inc., Center Valley, USA). Calcofluor was excited by a 400-410 BandPass filter and detected via a 455 nm LongPass filter.
- 146 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
2.5.2.5. Confocal laser scanning microscopy (CLSM)
CLSM was used to characterise the protein network distribution within pasta strands.
Sample preparation: 15 µm thick cross sections were stained with 0.01% (w/v) Fuchsine Acid in 1% (v/v) acetic acid, as described by Fardet et al (1998).
Image acquisition and processing: images were acquired using an inverted Confocal Laser Scanning Microscope (CLSM) (Zeiss Axiovert 200M) with attached Zeiss LSM510 META imaging system (Carl Zeiss, Jena, Germany) in the Montpellier RIO Imaging (MRI) facility, as described by Petitot et al (2009c). For each cross-section, images were acquired at 3 locations within pasta: external, intermediate and central region, without any overlapping between the images. For each of the 3 pasta samples (durum wheat, split pea or faba bean pasta), 3 strands were cut and 3 sections per strand were observed, totalling to 81 images. After a normalisation procedure, the protein network was characterized by analysing the grey level granulometry based on mathematical morphology (Devaux, Robert, Melcion & deMonredon, 1997; Rouille, Della Valle, Devaux, Marion & Dubreil, 2005). Erosion/dilation curves were analysed by principal component analysis (PCA). Similarity map could be drawn from principal component (PC) scores, whereas the loading analysis revealed the weight of each erosion or dilation step in their computation. ANOVA was performed on PC scores to supplement PCA observations. Image analysis was performed with Matlab (v7.0.4) software (The MathWorks, Paris, France) using dedicated toolboxes (image processing and PLS toolbox v3.5 (Eigenvector Research Inc., Manson USA).
2.6. Protein extractability during pasta processing
2.6.1. Protein extraction procedure
Proteins were extracted in triplicate from semolina, split pea and faba bean flours, freezedried mixed dough, freeze-dried extruded dough, dried pasta and freeze-dried cooked pasta, according to a modified method of Morel, Dehlon, Autran, Leygue & Bar-L'Helgouac'h (2000), as described by Petitot et al (2009c). The extraction procedure consists of two successive extractions. The first extraction is conducted at 60°C for 80 min with a sodium phosphate buffer containing 1% of sodium dodecyl sulphate (SDS, 0.1M) to extract SDS-soluble proteins. Then - 147 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
the pellet is suspended at 60°C for 60 min with 5ml of the SDS-phosphate buffer containing 20mM of dithioerythritol (DTE) and sonicated for 5min in order to extract SDS-insoluble proteins, referred as DTE-soluble proteins. The fraction that was extracted neither in SDS nor in DTE followed by sonication constituted the insoluble fraction.
2.6.2. Protein size distribution measurement
The protein size distribution of cooked pasta was studied by Size-Exclusion High Performance Liquid Chromatography (SE-HPLC). The SE-HPLC apparatus (Waters model LC Module1 plus) was equipped with an analytical column TSK G4000-SW (Merck, France) (7.5 × 300 mm) and a guard column TSK G3000-SW (Merck, France) (7.5 × 75 mm), as previously described (Morel et al., 2000). Once corrected for their different solid-to-solvent ratios during extractions, areas (in arbitrary units) of SDS-soluble and DTE-soluble proteins were added and the sum (i.e. total extractable proteins) was expressed as percents of the corresponding total area calculated for semolina (for durum wheat pasta) or a blend of 65% semolina and 35% legume flour (for legume pasta) (on equivalent dry protein basis). Each SE-HPLC profile of SDS-soluble proteins was arbitrary divided into five fractions (S1 to S5). Apparent molecular weights were estimated by calibrating the column with protein standards according to Redl, Morel, Bonicel, Vergnes & Guilbert (1999). Fraction S1 corresponded to polymeric proteins eluted at the void volume of the column (blue dextran, Mr = 2000 kDa). Fraction S2 corresponded to proteins ranging from Mr ≈ 780 to 95 kDa. Fractions S3 and S4 corresponded to proteins ranging from Mr ≈ 95 to 52 kDa and from 52 to 21 kDa respectively. Fraction S5 corresponded to the smallest monomeric proteins (Mr < 21 kDa). The second extract, obtained after solubilisation by the combined action of DTE and sonication, characterised SDS-insoluble proteins whose molecular weight exceeded 2000 kDa before sonication and solubilisation in DTE. Data were subjected to analysis of variance (ANOVA) followed by the Fisher’s least significant difference (LSD) test to compare means at the 5% significance level by using Microsoft Xlstat software 2008 (Addinsoft, Paris, France).
2.7. DSC analysis of starch in dry and cooked spaghetti
Differential scanning calorimetry (DSC) measurements were conducted on a DSC Q200 modulated (TA Instruments, New Castle, USA) calibrated with indium and an empty pan as a reference. Durum wheat semolina, dried pasta and freeze-dried cooked pasta samples were - 148 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
ground and sieved to pass through 250µm mesh screen. Ground samples were accurately weighed in aluminium hermetic pans. Water was added with a micropipette with a water-to-solid ratio of 4. Then the pans were sealed, reweighed and allowed to equilibrate for 1h at room temperature. The analyses were performed from 10 to 120°C at a heating rate of 10°C/min using an empty pan as a reference. For each endotherm, the onset (To), peak (Tp) and conclusion (Tc) temperatures, and the gelatinization enthalpy (∆H) were computed by using the TA instruments analysis software program. Temperature ranges (∆Tr = Tc-To) were calculated. Each experiment was repeated four times for each sample. Data were subjected to analysis of variance (ANOVA) followed by the Fisher’s least significant difference (LSD) test to compare means at the 5% significance level by using Microsoft Xlstat software 2008 (Addinsoft, Paris, France).
2.8. In vitro starch digestibility of cooked spaghetti
An in vitro starch digestion method that mimics human digestion was performed by Englyst Carbohydrates Ltd. (Englyst et al, 1999). This method is based on the measurement by HPLC of the glucose released from a test food during timed incubation with digestive enzymes under standardized conditions. The glucose released from starch within 20 min incubation corresponds to the rapidly available glucose (RAG). It has been found to be a strong predictor of postprandial glycemia (Englyst et al, 1999).
3. Results 3.1. Chemical composition, solubility in water and heat coagulation of soluble proteins in durum wheat semolina and legume flours Durum wheat semolina contains two major components, starch (~78%) and proteins (~13%), and minor components such as lipids and fibres (Table 5.1). Gliadins (alcohol-soluble) and glutenins (acide/alkali soluble) represent 80% of total proteins vs. 20% for albumins (watersoluble) and globulins (salt-soluble) (Osborne, 1907). Glutenins are composed of high molecular weight (HMW) subunits linked by disulphide bonds (Schofield, 1986). They can form both intra and inter-molecular bonds whereas gliadins can form only intramolecular ones (Kokini et al., 1994; Singh et al., 2004). Compared to durum wheat semolina, split pea and faba bean flours contain lower amounts of starch (48 and 44%) and higher amounts of proteins (21.4 and 29%), fibres (13.4 and 7.3 %, mainly insoluble) and lipids (2.5 and 2.2 %) (Table 5.1). Legume proteins
- 149 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
can be fractioned into water-soluble albumins, salt-soluble globulins, alcohol soluble prolamins and acid/alkali-soluble glutelins. The proportion of the different fractions was reported in the literature and depending on the cultivars and the extraction procedure used, different results were obtained. However, globulins and albumins remained the main protein fractions of split pea and faba bean. Globulins and albumins represented 45-60% and 14-30% of split pea proteins; 5575% and 2-33% of faba bean proteins respectively (El-Fiel et al., 2002; Leterme et al., 1990; Pasqualini et al., 1991; Sell, Steinhart & Paschke, 2005; Zheng, Fasina, Sosulski & Tyler, 1998).
Table 5.1. Chemical composition, protein solubility (in water, pH 6,3) and heat coagulation of soluble proteins at 100°C of durum wheat (DW) semolina, split pea (SP) flour, faba bean (FB) flour and blends of 35% legume flour and 65% durum wheat semolina. Protein content
Protein solubility in water
(% db)
(% total proteins)
DW semolina
13.3 ± 0.2
11.1
SP flour
21.4 ± 0.4
FB flour
29.0 ± 0.8
Heat coagulation of soluble proteins (% total soluble proteins)
Fibres Starch
Lipids Total
Soluble
(% db)
(% db)
(% db)
(% db)
70.5 ± 2.8
77.6 ± 0.3
1.7 ± 0.04
2.4
0.7
64.4
66.6 ± 4.5
47.9 ± 0.5
2.5 ± 0.09
13.4
1.1
78.8
45.5 ± 1.4
44.4 ± 0.4
2.2 ± 0.09
7.3
1.0
35% SP flour + 65% DW semolina a
16.1
67.0
1.9
6.2
0.8
35% FB flour + 65% DW semolina a
18.8
66.0
1.8
4.1
0.8
a
results obtained by calculation
Durum wheat semolina proteins were characterised by a low solubility in water (11,1%) (Table 5.1). In comparison, legume flour proteins were highly soluble, especially faba bean flour (78.8%) which is in accordance with Carbonaro et al. (1997). Heating of soluble proteins caused a decrease in protein solubility due to protein denaturation and coagulation (Klepacka, Porzucek & Kluczynska, 1997). 70 and 65% of soluble proteins coagulated during cooking for durum wheat and split pea vs. 45% for faba bean flour (Table 5.1). Faba bean proteins were less sensitive to thermal treatment than split pea proteins. In accordance with our results, (Klepacka et al., 1997) found also that pea was more affected by heat treatment than were bean and lupin proteins. - 150 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
3.2. Porosity of dry spaghetti
Dry durum wheat pasta presented a low porosity (5.9%) with a major fraction (90%) of mezopores (2 – 50 nm) and a minor fraction (10%) of macropores (0.05 – 2 µm). Pagani et al. (1989) reported a slightly lower porosity (4%) in dry pasta that could be due to the use of a Teflon die during pasta shaping (instead of a bronze die in our study) (Lucisano, Pagani, Mariotti & Locatelli, 2008). The addition of 35% of split pea or faba bean flour to semolina did not have a noticeable impact on the total porosity of dry pasta. However, it tended to decrease the proportion of mezopores (80%) in favour of macropores (20%), probably as a result of the higher content in insoluble fibres.
3.3. Water absorption kinetic during cooking
In durum wheat pasta and faba bean pasta, water was continuously absorbed during cooking and overcooking (Figure 5.1), as already observed by Baiano et al. (2006). Split pea and faba bean pasta presented a lower optimal cooking time (OCT) (Table 5.2) and lower hydration rate (Figure 5.1) than durum wheat pasta. As a consequence, the ratio between the increase in weight and the weight of dry spaghetti (∆W/W0) at optimal cooking time decreased from 1.28 for durum wheat pasta to 1.16 for both legume pasta (Table 5.2). Chillo et al. (2008a) also found a lower hydration rate of pasta made from amaranthus (85%, db) supplemented with quinoa, broadbean or chickpea flour (15%, db) compared to durum wheat spaghetti that was also accompanied by a lower optimal cooking time. Lower water uptake was also reported in pasta fortified with 15% of chickpea flour or 20% of green pea flour and cooked at their optimal cooking time (Sabanis et al., 2006; Zhao et al., 2005).
- 151 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
Figure 5.1. (Wt-W0)/Wt plotted as a function of cooking time illustrating water absorption kinetic for durum wheat (DW) pasta, split pea (SP) pasta and faba bean (FB) pasta. Experimental data and best fit (). OCT: optimal cooking time
Table 5.2. Water absorption and in vitro starch digestibility (RAG and RS, on a dry basis) of durum wheat (DW) pasta, split pea (SP) pasta and faba bean (FB) pasta cooked at their optimal cooking time (OCT) + 1min. OCT + 1 min (min)
(Wt-W0)/W0
Rapidly available glucose (%)*
Resistant starch (%)
DW pasta
10.3 ± 0.3 a
1.28 ± 0.01 a
62.5 ± 1.0 a
1.04 ± 0.16 b
SP pasta
9.6 ± 0.1 b
1.16 ± 0.01 b
59.4 ± 1.1a
1.69 ± 0.25 a
FB pasta
9.6 ± 0.1 b
1.16 ± 0.00 b
59.8 ± 2.8 a
1.68 ± 0.07 a
OCT: optimal cooking time. *: % of available carbohydrates (Wt-W0)/W0 : ratio between the change in weight at the OCT+1 min and the initial weight. Means ± SD with the same superscript within a column are notr significantly different (p > 0.05)
- 152 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
3.4. Microstructure 3.4.1. Environmental electronic scanning microscopy (ESEM) and particle size distribution of durum wheat and legume starch granules. ESEM and particle size distribution were used to characterise starch granule shape and size (Figure 5.2). According to ESEM observations, durum wheat starch (Figure 5.2 A) was composed of large lenticular and small spherical starch granules, in accordance with Buléon et al. (1998). In split pea and faba bean (Figure 5.2 B and C), large oval or irregularly shaped starch granules could be observed, in accordance with Colonna, Buléon & Mercier (1981). The analysis of particle size revealed a bimodal distribution of wheat starch granules with a major fraction (80%) of large starch granules (average diameter 25-30µm) and a minor fraction (20%) of smaller ones (average diameter 3-4 µm) (Figure 5.2 D), in accordance with ESEM images. Legume starch granules presented also a bimodal distribution but with a higher proportion (90%) of larger starch granules (average diameter 25µm) and a lower proportion (10%) of smaller ones (average diameter 3-4µm) that were more difficult to distinguish on ESEM images.
Figure 5.2. ESEM images (A, B, C) and particle size distribution (D) of starch granules isolated from durum wheat semolina (A), split pea flour (B) and faba bean flour (C).
- 153 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
3.4.2. Bright field light microscopy
Figure 5.3 shows representative views of cooked (OCT + 1min) durum wheat and legume pasta, from the core to the external region of pasta, from left to right. Starch granules appeared to be surrounded by a continuous protein network. For durum wheat and legume pasta, gradual structural changes were visible from the core to the external region of pasta, as already observed by Cunin et al (1995) and Heneen et al (2003). In the central and intermediate regions, starch granules presented a limited degree of swelling whereas in the external region, starch granules were largely deformed, swollen and it was difficult to differentiate them from proteins. As previously seen on native starch granules, it was possible to distinguish starch coming from durum wheat or legume flour, in the central and intermediate regions, according to their shapes. Durum wheat starch were mainly characterised by an elongated oval shape whereas legume starch granules appeared rather round or oval. Legume starch granules were randomly distributed within pasta strand, showing no specific organization.
Figure 5.3. Bright field light microscopy images of the internal structure of cooked (OCT+1min) pasta from the central (left) to the external region (right) of A) durum wheat pasta, B) split pea pasta and C) faba bean pasta. Proteins were stained with Fast Green and starch granules with iodine. The proteins stained green-turquoise while starch granules stained blue-violet.
- 154 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
3.4.3. Polarised light microscopy Cross sections of spaghetti cooked at their optimal cooking time (OCT) + 1min observed under polarised light are presented Figure 5.4. Ungelatinised native starch granules exhibit birefringence (or a “Maltese cross” pattern) under polarised light. Total loss of birefringence in durum wheat pasta (Figure 5.4 A) indicates that starch granules were completely gelatinised. In contrast, in legume pasta, few starch granules maintained their birefringence in the core of pasta strand (Figures 5.4 B and C). Analysis of birefringence was also conducted on legume pasta cooked at OCT + 3 min (i.e. 11.6 min), which corresponds to the time necessary to obtain the same water absorption (∆W/W0) than durum wheat pasta cooked at OCT + 1 min (i.e. 10.3 min). At this cooking time, some starch granules still exhibited birefringence in the core of legume pasta strand, confirming that some starch granules are resistant to gelatinisation (results not shown). This could reflect a lack of water in the core of pasta strand and / or a higher crystallinity of legume starch granules.
Figure 5.4. Polarised light microscopy images of the central region (A, B, C) and of the whole cross section (D) of cooked (OCT +1min) pasta. A) Durum wheat pasta, B) Split pea pasta, C) Faba bean pasta and D) overview of split pea pasta strand. Starch granules exhibiting birefringence with the typical Maltese cross pattern appear in white. - 155 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
3.4.4. Fluorescence microscopy Figure 5.5 shows fibres of durum wheat and legume pasta. Durum wheat pasta was characterised by the presence of small fibre fragments (~40µm) and few larger ones (200µm), randomly distributed (Figure 5.5 A). In comparison, split pea and faba bean pasta were characterised by higher amounts of randomly distributed fibres (Figures 5.5 B and C) coming from legume flour and durum wheat semolina. In faba bean pasta, fragments were of the same range of size than in durum wheat pasta, whereas in split pea pasta some large cellular structures and large fibre fragments were visible (Figure 5.5 D) although legume flours were dehulled. Aleurone and pericarp structures coming from durum wheat was also visible in durum wheat and legume pasta (Figure 5.5 E).
Figure 5.5. BFLM of cooked (OCT + 1min) durum wheat pasta (A), split pea pasta (B), faba bean pasta (C), and fibre fragments in split pea pasta (D) and durum wheat pasta (E) showing fibres stained by Calcofluor. Al. : Aleurone, Per : Pericarp, Cell : Cellular structure.
- 156 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
3.4.5. Confocal laser scanning microscopy (CLSM)
Examples of CLSM images of cooked (OCT+1min) durum wheat and legume pasta are presented in Figure 5.6. The protein network appeared in white (fluorescent area), whereas non fluorescent (dark) area could be assimilated to starch granules (Figures 5.6 A, B, C). The corresponding erosion-dilation curves are also presented (Figure 5.6 D). The left side of the curve, corresponding to dilations, and the right side corresponding to erosions, give information about dark cell (starch granule) size and protein film thickness, respectively.
Figure 5.6. CLSM images (A, B, C) and corresponding erosion-dilation curves (D) of the external region of cooked (OCT + 1min) durum wheat pasta (A), split pea pasta (B) and faba bean pasta (C). 81 images were acquired in total for durum wheat and legume pasta and their erosiondilation curves were processed by principal component analysis (PCA) and ANOVA was performed on principal component scores (Figure 5.7 and Table 5.3). The first two principal components accounted for 96.7% of the total variance (76.4 and 20.3% for PC1 and PC2 - 157 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
respectively) (Figure 5.7 A). Images from the external regions of pasta strands were mainly found on the negative side of PC1 while images from the intermediate and central regions were found on the positive side. According to loading 1 (Figure 5.7 B), external regions were characterised by a higher amount of thin protein films and a higher amount of large dark areas that could be related to a higher starch swelling. Moreover, whatever the location, split pea pasta had lower PC1 scores compared to durum wheat and faba bean pasta, and was characterised by a higher amount of large starch granules and higher amount of thin protein films. ANOVA revealed a significant impact of the location and of the raw material with no significant interaction between both on PC1 scores (Table 5.3). Whatever the pasta considered, images coming from the external, intermediate and central regions were significantly different. Whatever the location, split pea pasta were found significantly different from durum wheat and faba bean pasta. Analysis of principal component 2 and ANOVA on PC2 scores revealed there was an influence of the raw material, with a significant interaction with the location on PC2 scores (Table 5.3). Split pea images were mainly found on the positive side of PC2 while faba bean images were mainly found on the negative side, durum wheat images being situated between both (Figure 5.7 A). According to loading 2, split pea pasta was mainly characterised by a higher amount of thin protein films and lower amount of large starch granules compared to faba bean pasta. Table 5.3. Two way analysis of variance and subsequent LSD test of the scores on principal component 1 (PC1) and principal component 2 (PC2) of the principal component analysis of the erosion-dilation curves for durum wheat (DW) pasta, split pea (SP) pasta and faba bean (FB) pasta Effect of raw material ∗
PC1
PC2
Effects ∗
df
p-value
DW
SP
FB
A: Raw material
2
< 0.0001
0.416 a
-0.776b
0.333 a
B: Location
2
< 0.0001
A*B
4
0.588
A: Raw material
2
0.005
B: Location
2
0.083
A*B
4
0.031
-0.051
0.529
∗
Effect of location ∗ Central
InterExternal mediate
1.907 a
1.021 b
-3.004 c
0.325
0.023
-0.355
-0.486
Main effects (raw material, location) and their interactions were first included in the model then non-significant interactions (P >0.05) were excluded
- 158 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
Figure 5.7. (A): Principal component analysis of the granulometric curves of cooked durum wheat pasta (DW), split pea pasta (SP) and faba bean pasta (FB), in the central (c), intermediate (i) and external regions (e). (B): loadings 1 and 2 corresponding to principal components 1 and 2.
- 159 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
3.5. Protein reticulation during processing of durum wheat, split pea and faba bean pasta
The protein size distribution of semolina and blends of 35% legume flour and 65% semolina was analysed after each pasta processing step by SE-HPLC. Proteins were first extracted with a phosphate buffer containing sodium dodecyl sulphate (SDS) to disrupt the electrostatic, hydrophobic and hydrophilic interactions. The SDS-insoluble extract was then treated with dithioerythritol (DTE) and sonicated to disrupt the covalent disulphide bonds and was referred as the DTE-soluble fraction. The fraction being not soluble in SDS nor in DTE combined to sonication was considered as the unextractable protein fraction.
3.5.1. Characterisation of durum wheat semolina and durum wheat semolina-legume flour blends
80% of durum wheat semolina proteins were soluble in SDS (i.e. proteins linked by noncovalent interactions) (Figire 5.8 A). The addition of 35% split pea (SP) or faba bean (FB) flour to durum wheat semolina increased the protein solubility in SDS to 87% and 89% respectively (Figures 5.8 C and E). An example of the SE-HPLC profile of the SDS soluble extract of durum wheat semolina and wheat-legume blends is shown Figure 5.9. The SDS soluble extract of each sample was arbitrary divided into five major fractions (S1 to S5) and the relative proportion of each fraction was calculated (Figures 5.8 B, D, F). In durum wheat semolina, fractions S1 (6%) and S2 (19%) could be assimilated to polymeric glutenin proteins, fractions S3 (7%) and S4 (34%) could be assimilated to gliadins and fraction S5 (14%) corresponded to water-soluble proteins (albumins and globulins) (Morel et al, 2000) (Figures 5.8 B and 5.9 A). Split pea and faba bean proteins were mainly composed of proteins with a molecular weight inferior to 100 kDa (Figure 5.9 A). As a consequence, the addition of 35% legume flour to durum wheat semolina changed the SEHPLC profile of the SDS-soluble fraction (Figure 5.9 B). In blends of semolina and split pea flour and blends of semolina and faba bean flour, fractions S1 (4% for both) and S2 (16 and 15% respectively) could be mainly attributed to durum wheat proteins, fraction S3 (15 and 23% respectively) characterised legume proteins, fraction S4 (34 and 33% respectively) contained a major fraction of durum wheat proteins and a minor fraction of legume proteins and fraction S5 (17 and 15% respectively) contained both proteins in equivalent proportions (Figures 5.8 D and F).
- 160 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
Figure 5.8. Peak areas of SE-HPLC elution profiles of SDS-soluble, DTE-soluble and unextractable protein fractions (A, C, E). Peak surfaces of each SDS-soluble fraction (B, D, F). (A, B): durum wheat pasta, (C, D) split pea pasta and (E, F): faba bean pasta. Bars bearing different letters are significantly different from each other (p< 0.05). DW: durum wheat pasta, SP: split pea pasta, FB: faba bean pasta.
- 161 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
Figure 5.9. SE-HPLC profiles of SDS-soluble proteins extracted from the raw materials (A) and from blends used to produce pasta (B). Profiles were adjusted for protein content. SEHPLC profiles were divided into 5 major fractions from S1 to S5. DW: durum wheat; SP: split pea; FB: faba bean.
- 162 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
3.5.2. Impact of mixing, extrusion, drying and cooking on protein size distribution
Protein solubility during processing of durum wheat pasta is presented in Figures 5.8 A and B. Mixing induced an increase by 9% in the SDS-solubility that was mainly due to an increase in the S1 and S2 fractions. This was accompanied by a decrease in the DTE-soluble fraction. This indicates that mixing induced a depolymerisation of larger and smaller glutenin polymers (S1 and S2), probably as a result of some disulphide bonds breakage. Extrusion and drying led to a loss in SDS-solubility by 4 and 15% respectively, which could be attributed to a decrease in the S1 and S2 fractions. This decrease in the SDS-soluble fraction was counterbalanced by an increase in the DTE-soluble fraction. Extrusion and drying promoted the aggregation of glutenin polymers, probably through disulphide bonds, leading to the formation of a larger protein aggregate that became insoluble in SDS. Finally, cooking led to a marked decrease in the SDSsolubility of all protein fractions (from S1 to S5). This was accompanied by an increase in the DTE-soluble fraction (by 45%) and the formation of unextractable protein aggregates (12%). Cooking induced a major protein aggregation, involving the formation of disulphide bonds but also other covalent bonds.
Protein solubility during processing of split pea pasta and faba bean pasta is presented Figures 5.8 C, D and 5.8 E, F respectively. Compared to durum wheat pasta, legume pasta had a higher proportion of SDS-soluble proteins and a lower proportion of DTE-soluble proteins, whatever the processing step considered, indicating a higher proportion of proteins linked by non-covalent interactions and a lower proportion of proteins linked by disulphide bonds (Figures 5.8 C and E). Moreover, after cooking, a lower proportion of unextractable proteins could also be observed, especially in faba bean pasta, indicating a lower proportion of proteins linked by strong covalent bonds. The higher proportion of SDS-soluble proteins in legume pasta was mainly due to a higher proportion of S3 fraction, i.e. albumins and / or globulins, prolamin, glutelin, coming from legume flour. Moreover, in contrast to durum wheat pasta, legume pasta differed in the evolution of the S1 and S2 fractions during processing. In both legume pasta, S1 was not affected by mixing and extrusion and S2 increased during extrusion. The reactivity of larger and smaller glutenin polymers appeared to be affected by the presence of legume proteins. However, further investigation would be necessary to determine in more details the type of interaction involved between the larger glutenin polymers and legume proteins during mixing and extrusion steps.
- 163 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
3.6. DSC analysis in dry and cooked spaghetti
Results of DSC analyses are presented in Table 5.4. In durum wheat semolina, legumewheat blends and dry pasta, two endothermic transitions were observed corresponding to the gelatinisation of starch in excess water (at approximately 65°C) and the reversible dissociation of amylose-lipid complexes (at approximately 95°C) in accordance with Yue et al. (1999) and Zweifel et al. (2000). In cooked pasta, no peak corresponding to the first endothermic transition could be observed, confirming that starch was totally gelatinised. The second endothermic transition was not analysed in cooked pasta as the characteristics of the melting of amylose-lipid complexes should not change after cooking, as previously observed by Cunin et al. (1995).
Table 5.4. DSC measurements obtained from the first and the second endothermic peak of DSC on durum wheat (DW) semolina, blends of 35% split pea (SP) or faba bean (FB) flour and 65% DW semolina, and dried spaghetti. Means (n=3) with the same superscript within a column are not significantly different (P > 0.05). To: onset temperature; Tp: peak temperature; Tc: conclusion temperature, ∆Tr = Tc-To: temperature range, ∆H: enthalpy. Second endothermic transition (amylose-lipid complexe dissociation)
First endothermic transition (starch gelatinisation) To (°C)
DW semolina and legume flours
Dried spaghetti
Tp (°C)
Tc (°C)
∆Tr (°C)
∆H (J/g db)
To (°C)
Tp (°C)
∆H (J/g db)
DW semolina
54.6
c
62.5 c
70.3 b
15.7 b
5.8 a
83.8 a
93.4 ab
1.5 ab
35% SP flour + 65% DW semolina
56.3 a
64.2 a
78.1 a
21.8 a
5.4 b
84.2 a
94.1 a
1.3bc
35% FB flour + 65% DW semolina
56.2 a
63.6 ab
77.3 a
21.1 a
5.1 c
84.7 a
94.1 a
1.7 a
DW pasta
54.0 b
63.1 bc
70.0 b
16.0 b
5.0 c
84.2 a
91.9 a
1.5 ab
SP pasta
56.1 a
63.9 ab
78.4 a
22.3 a
4.5 d
84.2 a
94.1 a
1.1 c
FB pasta
56.0 a
63.5 ab
77.2 a
21.2 a
4.5 d
85.9 a
93.7 ab
1.5 ab
- 164 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
The addition of 35% legume flour to durum wheat semolina increased the onset and peak temperatures of the first endothermic transition, probably as a result of the higher transition temperatures of legume starch, especially of pea starches (Colonna et al., 1979). Indeed, high transition temperatures have been reported to result from a high degree of crystallinity, which provides structural stability and makes the granules more resistant towards gelatinization (Singh, Singh, Kaur, Sodhi & Gill, 2003). The different transition temperatures between legume and durum wheat starches could also explain the higher temperature range of legume-wheat blends. The lower gelatinisation enthalpy observed in legume-wheat blends could be due to lower starch contents of legume flours. The addition of legume flour had no significant impact on the dissociation of amylose-lipid complexes. Durum wheat and legume pasta presented a lower gelatinisation enthalpy than the raw materials used to produce pasta. This is in accordance with previous studies on durum wheat pasta (Petitot et al., 2009c; Yue et al., 1999; Zweifel et al., 2000). Starch damage during extrusion or gelatinisation of some starch granules during drying, which require less energy to melt, could explain this result (Biliaderis, 1990). The impact of legume flour addition to durum wheat semolina on the onset temperature, temperature range and gelatinisation enthalpy was still detected in dry pasta. However, no significant difference could be observed on the peak temperature of the first endothermic transition among dry pasta samples. Concerning the second endothermic transition, split pea pasta were characterized by a lower enthalpy compared to durum wheat and faba bean pasta, that could reflect a lower amount of amylose-lipid complexes.
3.7. Starch digestibility
The physiological effects of starch depend on the rate an extent of digestion in the small intestine, which strongly influences the glycaemic response to different foods. The glycaemic index (GI) is calculated as the measured glycaemic response to a portion of a test food that contains 50 g “available” carbohydrate (which includes fructose) expressed as a percentage of the glycaemic response to the same amount of available carbohydrate from a standard food (glucose or white bread) eaten by the same subject (Jenkins, 1981). The in vitro technique used here determined the rapidly available glucose (RAG) value by measuring the amount of glucose released during timed incubation with digestive enzymes. It was expressed in g for 100g available carbohydrates (total carbohydrates minus resistant starch). The RAG value represents the amount of glucose that would be expected to be rapidly available for absorption after eating, and it shows a positive correlation with the glycaemic index (Englyst et al, 1999). RAG values of - 165 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
cooked pasta are presented in Table 5.2. Durum wheat spaghetti obtained a low RAG value of 62.5 (g / 100 g available carbohydrates), which is in accordance with Englyst, Englyst, Hudson, Cole & Cummings (1999). The introduction of high quantity (35%, db) of split pea or faba bean flour did not have any significant impact on the RAG value of pasta. Based on RAG values, the post-prandial glycaemic response would be reached slowly after consumption of either durum wheat pasta or legume pasta. Moreover, legume pasta contained high amounts of resistant starch (1.7% for split pea pasta and faba bean pasta vs. 1.04 % for durum wheat pasta) (Table 5.2) that would be fermented by colonic bacteria to short-chain fatty acid and would have therefore beneficial physiological effects on the colonic health (Cummings & Englyst, 1995; Tharanathan et al., 2003).
4. Discussion and conclusion The objectives of this study were to investigate the impact of durum wheat semolina substitution by split pea or faba bean flour on pasta structure, studied at multi-scales, and the inherent consequences on its starch digestibility. High level (35%) of legume flour substitution was used in order to increase the nutritional composition of pasta (enrichment in proteins, fibres, vitamins, minerals and amino-acid complementarity).
Cooked durum wheat pasta was characterised by a protein network entrapping starch granules, in accordance with Fardet et al. (1998b) and Heneen et al. (2003). Structural microscopic changes from the external region to the core of pasta could be observed, as a consequence of the moisture gradient inside cooked pasta strand (Horigane et al., 2006). In the external region, starch granules were fully swollen and it was difficult to differentiate them from proteins in accordance with Heneen et al. (2003). The protein network was characterised by higher amounts of thin protein films in the external region compared to intermediate and central regions. In the central region, starch granules were partly swollen but fully gelatinised as confirmed by DSC measurements and polarised light microscopy. At a lower scale, the protein network in cooked pasta was highly aggregated with a major fraction of proteins (70%) linked by disulphide bonds and a minor fraction of proteins (12%) linked by other covalent bonds. The study of starch digestibility, i.e. the measurement of rapidly available glucose (RAG) value revealed that durum wheat pasta had a low RAG value (62.5 g for 100 g available carbohydrates). This slow glucose release can be ascribed to the compact structure of pasta (low porosity) and the presence of an aggregated protein network entrapping starch granules, which - 166 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
may delay the degradation of the protein network by proteases and thus delay the accessibility of starch to α-amylases.
The substitution of durum wheat semolina with 35% split pea or faba bean flour induced some structural changes to pasta. At a microscopic level, cooked legume pasta was also characterised by a protein network entrapping starch granules with some structural changes from the external region to the core of pasta. In contrast to durum wheat pasta, higher amounts of randomly distributed fibres could be observed. In particular, split pea pasta presented big cellular structures and fibre fragments. The introduction of fibre fragments in pasta may weaken its structure, due to the disruption of the protein network, as previously hypothesised by Tudorica et al. (2002) in pasta fortified with 7.5% of pea fibres. However, according to our microscopic observations, no disruption of the protein network was apparent. In faba bean pasta, the distribution of the protein network was similar to the one in durum wheat pasta. The higher protein content of both legume pasta did not induce a thicker protein network. Split pea pasta presented even higher amounts of thinner protein films compared to faba bean and durum wheat pasta, whatever the location in pasta strand. At a macromolecular scale, cooked legume pasta and especially faba bean pasta presented a higher proportion of proteins linked by non-covalent interactions and a lower proportion of proteins linked by disulphide and other strong covalent bonds, compared to durum wheat pasta. Legume proteins, mainly composed of globulins and albumins (El-Fiel et al., 2002; Leterme et al., 1990; Pasqualini et al., 1991; Sell et al., 2005; Zheng et al., 1998) are poor in sulphur-containing amino-acids, especially cystein residues (Alonso et al., 2000; Carbonaro et al., 1997). The lack of sulphur groups could explain the lower potential of legume proteins to bind through disulphide bonds compared to gluten proteins (Li & Lee, 2000). Heat denaturation of legume protein would involve the exposure of hydrophobic residues, protein aggregation by intermolecular hydrophobic interaction and agglomeration of aggregates into a network structure (O'Kane, Happe, Vereijken, Gruppen & Van Boekel, 2004; Zheng et al., 1998).
The addition of legume flour had also an impact on the starch fraction. In the intermediate and central regions of pasta strand, legume starch granules could be differentiated from durum wheat starch by their different sizes and shapes. In contrast to durum wheat pasta, few starch granules exhibited birefringence under polarised light in the core of pasta strand, indicating the presence of ungelatinised starch granules. This could be a consequence of limited water content in the core of legume pasta strand. However, even after overcooking, few starch granules still - 167 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
exhibited birefringence in legume pasta showing that legume starch were more resistant to starch gelatinisation. The higher degree of crystallinity of legume starches may have provided structural stability, rendering the granules more resistant towards gelatinization (Singh et al., 2003). Moreover, a local competition for water between proteins, starch granules and fibres may have occurred. Protein denaturation and starch gelatinisation occur in the same range of temperature and moisture conditions (Cuq et al., 2003). They are competitive and antagonistic (Pagani et al., 1986) : the formation of the protein network is opposed to the swelling of the starch granules. Both transformations are controlled by water penetration which occurs concentrically toward the centre of spaghetti with cooking time (Horigane et al., 2006). The higher content in albumins, globulins and hydrophilic insoluble fibres in legume pasta may have impacted on the hydration kinetic of pasta. Insoluble fibres may have preferentially absorbed water inhibiting starch swelling and subsequent absorption of water. Different structural transformation kinetics of pasta components may have occurred. In the core of pasta strand, the formation of the protein network may have preceded the gelatinisation of the starch granules, preventing starch swelling and gelatinisation.
Some structural changes appeared to occur due to the fortification of pasta with split pea and faba bean flours. Some of them may have contributed to the increase or the decrease in starch digestibility. Indeed, previous work of Goni et al. (2003) showed that the addition of 25% chickpea flour to durum wheat flour induced a lower degree of in vitro starch hydrolysis and of the in vivo glycaemic index (GI). They attributed this lower GI to the higher amount of the indigestible fraction (non-starch polysaccharides, resistant starch, oligosaccharides) in chickpea pasta. Tudorica et al. (2002) found that pea fibre inclusion at 7.5% disrupted the protein network entrapping starch granules, leading to a higher in vitro enzyme susceptibility of starch granules. In our study, the addition of legume flour to durum wheat pasta did not have any significant impact on the in vitro digestibility of starch. Indeed, despite the slight differences in pasta structure observed in legume fortified pasta, starch digestibility remained as low as in durum wheat pasta. The higher content in fibres and in resistant starch may have contributed to a decrease in starch digestibility. In the other hand, the dilution of the gluten network by legume proteins and the presence of large fibre fragments (especially in split pea pasta) may have induced the formation of a weaker protein network. Both may have favoured the accessibility of enzymes to starch.
- 168 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
The substitution of durum wheat semolina with legume flour appeared to be an interesting way to increase the nutritional composition of pasta (protein, fibre, vitamin and mineral contents, amino acid complementarity) while keeping the low digestibility of its starch.
Acknowledgments This work was carried out with the financial support of the « ANR- Agence Nationale de la Recherche - The French National Research Agency » under the « Programme National de Recherche en Alimentation et nutrition humaine », project « ANR-05-PNRA-019, PASTALEG». The authors are very grateful to J. Bonicel (UMR IATE, Montpellier, France) for her technical assistance and to Martine Champ (INRA, UMR PhAN, Nantes, France), Isabelle Crenon (INRA/INSERM, UMR 476/1260) and Chantal Brossard (INRA, UR 1268 BIA, Nantes, France) for fibre, lipid and protein quantification, respectively. The authors are indebted to B. Chabi from the Montpellier RIO Imaging Platform (RIO Imaging facility, Montpellier, France) for imaging experiments and B. Vernus from DCC laboratory (UMR DCC, Montpellier, France) for microtome use. The authors wish to thank B. Bouchet, F. Guillon (UR BIA, Nantes, France) and G. Viennois (CNRS, UMR 386 BPMP, Montpellier, France) for helpful discussions about microscopic analyses.
- 169 -
- 170 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
Partie 2. Publication 6: Legume fortified pasta. Impact of drying and precooking treatments on pasta structure and inherent starch digestibility
Maud Petitot, Valérie Micard
A soumettre à Journal of Cereal Science
ABSTRACT The impact of technological treatments on the structure and in vitro starch digestibility (rapidly available glucose value) of legume fortified pasta was investigated. Legume fortified pasta was prepared from 65% of durum wheat semolina and 35% of split pea or faba bean flour. Four different technological treatments were applied: drying at low temperature 55°C (LT); drying with the application of a very high temperature 90°C at low moisture content (VHT.LM); lyophilisation; and precooking followed by LT-drying. Legume fortified pasta dried at LT served as reference. Lyophilisation induced a higher starch digestibility that could be attributed to the high porosity of pasta and the weakness of its protein network. In contrast, VHT.LM drying and precooking treatment led to a lower starch digestibility, probably as a result of the strengthening of the protein network at a macromolecular level, protecting starch from enzymatic attack.
Key-words: split pea – faba bean – durum wheat pasta – microstructure – gluten network
- 171 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
1. Introduction Pasta is a source of proteins and slow release carbohydrates (Granfeldt et al., 1991a; Jenkins et al., 1983). In pasta, starch is slowly digested and absorbed in the small intestine, leading to low postprandial blood glucose and insulin responses in humans (Björck et al., 2000). This low glycaemic index (GI) can be ascribed to its specific structure. The compactness of the structure (Bjorck et al., 1994; Wolever et al., 1986) and the encapsulation of starch granules by a continuous protein network (Colonna et al., 1990; Fardet et al., 1998b) are the main factors that can explain this low GI. Despite this interesting nutritional property, pasta is deficient in two of the essential amino-acids, lysine and threonin (Abdel-Aal et al., 2002; Kies et al., 1970). Non traditional ingredients, such as legume flours, have been thus incorporated into pasta for nutritional improvement (Bahnassey et al., 1986b; Goni et al., 2003; Zhao et al., 2005). Indeed, legumes represent a good source of proteins (rich in lysine), slow release carbohydrates, dietary fibres, vitamins and minerals (Carbonaro et al., 1997; de Almeida Costa et al., 2006; Tharanathan et al., 2003). However, modification of pasta composition through the incorporation of new raw materials can change its nutritional properties. Goni et al. (2003) demonstrated that fortification of pasta with 25% chickpea flour decreased in vitro and in vivo glycemic index of pasta. They attributed this change to the presence of non digestible constituents (resistant starch, oligosaccharides, polyphenols and lectins). Changes in pasta nutritional properties can also be attributed to changes in pasta structure due to the presence of new components. The inclusion of 7.5% pea fibres was shown for example to weaken pasta structure and induce higher starch digestibility (Tudorica et al., 2002). Besides the change in pasta formulation, the modification of process parameters can also be a way to change pasta structure and enhance or deteriorate its nutritional properties (Petitot et al., 2009a). For example, the application of a very high temperature (VHT) during drying of durum wheat pasta was shown to induce the formation of a strong protein network (De Zorzi et al., 2007; Holm et al., 1988b; Lamacchia et al., 2007; Petitot et al., 2009c) and decrease the in vitro digestibility of starch (Casiraghi et al., 1992) and proteins (De Zorzi et al., 2007; Petitot et al., 2009c). However, the impact of technological treatments, on pasta structure and consequence on its nutritional properties has not been investigated on legume fortified pasta yet. The objective of this study was therefore to determine the impact of high level of legume flour addition combined to the use of drying and precooking treatments on pasta structure, and the repercussions on starch digestibility. Four different treatments were thus applied on freshly extruded pasta affecting the structure of pasta at different scales (from macroscopic to molecular - 172 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
one), such as its compactness and the structure of both protein and starch components. Indeed, the compact structure of pasta (Barkeling et al., 1995; Wolever et al., 1986), the encapsulation of starch granules by a protein network (Colonna et al., 1990; Fardet et al., 1998b), and the physical structure of starch, such as its degree of gelatinisation were shown to affect the glycemic index (GI). A low temperature drying (LT); a very high temperature (VHT) drying with the VHTphase applied at low moisture content (VHT.LM); lyophilisation; and a precooking treatment followed by LT-drying were chosen to reach theses objectives. The LT-drying served as reference. Lyophilisation was used to change the porosity of pasta at a macro-mesoscopic scale. At a lower scale, the VHT.LM drying was used to strengthen the protein network as previously seen in durum wheat pasta (Petitot et al., 2009c). The precooking treatment followed by LT drying was applied in order to affect the structure of the starch fraction.
2. Materials and methods Durum wheat semolina was supplied by Panzani (Marseille, France). Dehulled split pea and faba bean flours were obtained from Terrena (Moisdon la Rivière, France).
2.1. Chemical composition of durum wheat semolina and legume flours
Total protein content was determined using Kjeldahl procedure with a nitrogen-to-protein conversion factor of 5.7. Total starch content was determined with an enzymatic assay kit (Megazyme, Co. Wicklow, Ireland). Fibres were determined by ISHA (Lonjumeau, France) according to the AOAC 985.29 method. Analyses were conducted in triplicate except for fibres (duplicate).
2.2. Pasta manufacturing
Pasta production: durum wheat spaghetti and spaghetti fortified with 35% split pea or 35% faba bean flour (db) were processed with a continuous pilot-scale pasta extruder (Bassano, Lyon, France) as described by Petitot et al. (2009e). Freshly extruded spaghetti was then submitted to one of the following treatments:
- 173 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
1) Drying in a pilot-scale drier (AFREM, Lyon, France) at low temperature 55°C (LT pasta) or at very high temperature 90°C with the application of the very high temperature at low moisture content (about 12%) (VHT.LM pasta) (Petitot et al., 2009c). 2) Drying in a pilot-scale lyophilizer (Lyofal SAS, Salon de Provence, France) (LYO pasta). 3) Precooking (120°C / 1.2 bars for 10 min) in a horizontal autoclave (B.B.C, Montpellier, France) and drying at low temperature 55°C (PreC pasta).
2.3 Porosity of dry pasta
The porosity and pore size distribution of dry pasta was determined on a ~3g sample at the Institut Européen des Membranes (UMR 5635, IEM, Montpellier) using an AutoPore IV 9500 mercury intrusion porosimeter (Micromeritics Instrument Corporation, Norcross, USA). The principle of this technique is that mercury behaves as a non-wetting liquid toward most substances. Mercury is forced to enter into the pores by applying a controlled increasing pressure (up to 400 MPa). The volume of mercury penetrated is detected by means of a capacitive system. Data of intruded volume of mercury versus applied pressure is obtained. The pressures are converted into pore sizes using Washburn equation: R= (-2γ cosθ)/P where γ is the surface tension of pure mercury (0.480 N.m -1); θ is contact angle between mercury and the solid (140°); P is the mercury pressure (Pa); and R is the pore radius (m). The total porosity was calculated as the total volume of intruded mercury at the maximum experimental pressure divided by the bulk volume of the sample. The distribution in macropores (> 50 nm) and mezopores (2-50 nm) was obtained from the curve of the cumulative intrusion volume vs. pore size.
2.4. Determination of optimal cooking time and hydration kinetic during cooking
- Optimal cooking time: dried spaghetti was cooked in Evian water containing 0.7% (w/v) of sodium chloride with a water-to-solid ratio of 20. Optimal cooking time (OCT) was indicated when the white core of the pasta disappeared when squeezed between two glass plates (Approved method 66-50 AACC 2000). All analyses on cooked pasta were made on pasta cooked at OCT + 1min. This cooking test was inappropriate for precooked pasta because of their high firmness. Precooked pasta was therefore cooked to have the same moisture content than LTdried pasta. - Hydration kinetic: water absorption tests were performed according to the method of del Nobile et al. (2005). Culture tubes containing 9 ml of distilled water were equilibrated at 100°C ± 1°C in - 174 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
a thermostated bath. 40 ± 0.5 mm spaghetti samples were immersed in the tubes, one for each tube. At given time (every 30 s for the first 10 min, every minute up to 16 min), samples were removed from the tubes, rapidly blotted and accurately weighed. 3 determinations were made for each sample. The ratio between the increase in weight and the weight of dry spaghetti (∆W/W0)= (W(t) – W0)/W0 was plotted as a function of time. Data were subjected to two-way analysis of variance (ANOVA) followed by the Fisher’s least significant difference (LSD) test to compare means at the 5% significance level by using Microsoft Xlstat software 2008 (Addinsoft, Paris, France).
2.5. Microstructure of cooked spaghetti
Polarised light microscopy and confocal laser scanning microscopy (CLSM) were used to characterise protein and starch fractions in split pea pasta.
2.5.1. Sample preparation
After cooking (OCT+1), water was decanted and pasta was cooled for 15 min before sectioning in a covered container at 25°C. Pasta was cut in small pieces and frozen quickly at 40°C in water plus a cryoprotector (OCT, Cellpath, Newtown, UK). Samples were cut at -20°C using a microtome (Microm HM 560, Walldorf, Germany). 8 µm thick cross sections were used for bright field, florescence and polarised light microscopy whereas 15 µm thick cross sections were used for confocal laser scanning microscopy.
2.5.2. Polarised light microscopy (PLM)
Loss of birefringence was detected under polarised light. Unstained images were observed using an Olympus BX61 microscope (Olympus America Inc., Center Valley, USA) equipped with a Hamamastu ORCA-AG camera (Hamamatsu Photonics K.K, Hamamatsu, Japan) and the Cell^P software (Olympus America Inc., Center Valley, USA).
2.5.3. Confocal laser scanning microscopy (CLSM):
CLSM was used to characterise the protein network distribution within pasta strands in LT, VHT.LM and PreC-pasta fortified with split pea flour. LYO pasta was observed but images were - 175 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
not included in the image analysis. Indeed, cutting pasta with the microtome induced the tearing of the protein network due to the fragility of LYO pasta structure. Sample preparation: 15 µm thick cross sections were stained with 0.01% (w/v) Fuchsine Acid in 1% (v/v) acetic acid, as described by Fardet et al (1998). Image acquisition and processing: images were acquired using an inverted confocal laser scanning microscope (CLSM) (Zeiss Axiovert 200M) with attached Zeiss LSM510 META imaging system (Carl Zeiss, Jena, Germany) in the Montpellier RIO imaging (MRI) facility, as described by Petitot et al (2009c). For each cross-section, images were acquired at 3 locations within pasta: external, intermediate and central region, without any overlapping between the images. For each of the 3 pasta samples (LT, VHT.LM and PreC-pasta), 3 strands were cut and 3 sections per strand were observed, totalling to 81 images. After a normalisation procedure, the protein network was characterized by analysing the grey level granulometry based on mathematical morphology (Devaux, Robert, Melcion & deMonredon, 1997; Rouille, Della Valle, Devaux, Marion & Dubreil, 2005). Erosion/dilation curves were analysed by principal component analysis (PCA). Similarity map could be drawn from principal component scores, whereas the loading analysis revealed the weight of each erosion or dilation step in their computation. Two-way ANOVA was performed on PC scores to supplement PCA observations. Image analysis was performed with Matlab (v7.0.4) software (The MathWorks, Paris, France) using dedicated toolboxes (image processing and PLS toolbox v3.5 (Eigenvector Research Inc., Manson USA).
2.6. Protein size distribution in dry and cooked spaghetti
2.6.1. Protein extraction procedure
Proteins were extracted in triplicate from semolina and legume flour blends, dried pasta and freeze-dried cooked pasta, according to a modified method of Morel et al. (2000), as described by Petitot et al. (2009c). This method is based on two successive extractions. The first extraction is conducted at 60°C for 80 min with a sodium phosphate buffer containing 1% of sodium dodecyl sulphate (SDS, 0.1M). Then the pellet is suspended at 60°C for 60 min with 5 ml of the SDS-phosphate buffer containing 20 mM of dithioerythritol (DTE) and sonicated for 5 min in order to extract SDS-insoluble proteins, referred as DTE-soluble proteins.
- 176 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
2.6.2. Protein size distribution measurement
The protein size distribution was studied by Size-Exclusion High Performance Liquid Chromatography (SE-HPLC). The SE-HPLC apparatus (Waters model LC Module1 plus) was equipped with an analytical column TSK G4000-SW (Merck, France) (7.5 × 300 mm) and a guard column TSK G3000-SW (Merck, France) (7.5 × 75 mm), as previously described by Morel et al. (2000). Once corrected for their different solid-to-solvent ratios during extractions, areas (in arbitrary units) of SDS-soluble and DTE-soluble proteins were added and the sum (i.e. total extractable proteins) was expressed as percents of the corresponding total area calculated for blends of 65% semolina and 35% split pea or faba bean flour (on equivalent dry protein basis). Apparent molecular weights were estimated by calibrating the column with protein standards according to Redl et al. (1999). The second extract was obtained after the combined action of DTE and sonication. Data were subjected to analysis of variance (ANOVA) followed by the Fisher’s least significant difference (LSD) test to compare means at the 5% significance level by using Microsoft Xlstat software 2008 (Addinsoft, Paris, France).
2.7. DSC analyses of dried and cooked spaghetti
Differential scanning calorimetry measurements were conducted on a DSC Q2000 (TA Instruments, New Castle, USA) calibrated with indium and an empty pan as a reference. Dried pasta and freeze-dried cooked pasta samples were ground and sieved to pass through 250 µm mesh screen. Ground samples were accurately weighed in aluminium hermetic pans. Water was added with a micropipette with a water-to-solid ratio of 4. Then the pans were sealed, reweighed and allowed to stand for 1h at room temperature. The analyses were performed from 10 to 120°C at a heating rate of 10°C/min using an empty pan as a reference. For each endotherm, the onset (To), peak (Tp) and conclusion (Tc) temperatures, and the gelatinization enthalpy (∆H) were computed by using the TA instruments analysis software program. Temperature ranges (∆Tr = Tc-To) were calculated. Each experiment was repeated three times for each sample. Data were subjected to two-way analysis of variance (ANOVA) followed by the Fisher’s least significant difference (LSD) test to compare means at the 5% significance level by using Microsoft Xlstat software 2008 (Addinsoft, Paris, France).
- 177 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
2.8. In vitro starch digestibility of cooked pasta An in vitro starch digestion method that mimics human digestion was performed by Englyst Carbohydrates Ltd. (Englyst et al, 1999). This method is based on the measurement by HPLC of the glucose released from a test food during timed incubation with digestive enzymes under standardized conditions. The glucose released from starch within 20 min incubation corresponds to the rapidly available glucose (RAG). It has been found to be a strong predictor of postprandial glycemia (Englyst et al, 1999). Data were subjected to two-way analysis of variance (ANOVA) followed by the Fisher’s least significant difference (LSD) test to compare means at the 5% significance level by using Microsoft Xlstat software 2008 (Addinsoft, Paris, France).
3. Results 3.1. Chemical composition Durum wheat semolina contains two major components, starch (~78%) and proteins (~13%), and minor components such as lipids and fibres (Table 5.5). In comparison, split pea and faba bean flours contain lower amounts of starch (48 and 44%) and higher amounts of proteins (21.4 and 29%), fibres (13.4 and 7.3 %) and lipids (2.5 and 2.2 %). Dried pasta containing 35% of split pea or faba bean flour should therefore contain 16-19% of proteins, 6766% of starch and 6-4% of fibres, respectively (Petitot et al., 2009b).
Table 5.5. Chemical composition of durum wheat (DW) semolina, split pea (SP) flour, faba bean (FB) flour and blends of 35% legume flour and 65% durum wheat semolina Fibres Proteins
Starch
Lipids
(% db)
(% db)
(% db)
Total (% db)
Durum wheat semolina
13.3 ± 0.2
77.6 ± 0.3
1.7 ± 0.04
2.4
0.7
Split pea flour
21.4 ± 0.4
47.9 ± 0.5
2.5 ± 0.09
13.4
1.1
Faba bean flour
29.0 ± 0.8
44.4 ± 0.4
2.2 ± 0.09
7.3
1.0
35% SP flour + 65% DW semolina
16.1
67.0
1.9
6.2
0.8
35% FB flour + 65% DW semolina
18.8
66.0
1.8
4.1
0.8
- 178 -
Soluble (% db)
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
3.2. Porosity of dry spaghetti Split pea and faba bean pasta dried at LT presented a low porosity (~5%), with a major fraction of mezopores (~80%) and a minor fraction of macropores (~20%). The application of a VHT.LM drying or a precooking treatment did not affect the total porosity of both legume pasta. In accordance with our results, Zweifel et al. (2001) demonstrated that the temperature applied during drying had not a great influence on the total porosity of pasta: the structure of pasta remained very dense and compact. However, we observed a small change in pore size distribution; i.e. a higher amount of mezoropores (90%) and lower amount of macropores (10%). In contrast, lyophilization increased the total porosity of pasta (14% and 14.9% for faba bean and split pea pasta, respectively), by increasing the amount of macropores (80%) at the expense of mezopores (20%). As expected and in accordance with Aguilera, Chiralt & Fito (2003) and Rhahman (2008), lyophilization resulted in a porous structure while hot-air drying resulted in a dense structure.
3.3. Optimal cooking time and spaghetti hydration kinetic during cooking
Whatever the treatment applied to pasta, split pea and faba bean pasta had similar optimal cooking time (OCT) + 1min (Table 5.6). For both split pea and faba bean pasta, PreC pasta was cooked for a longer time (11.7 min) and LYO pasta for a shorter time (7.6 min) compared to LT and VHT.LM pasta (9.6 min) (Table 5.6). The analysis of hydration kinetic during cooking revealed that three main groups could be distinguished: LT pasta; LYO pasta with a higher hydration kinetic; and a group formed by VHT.LM and PreC pasta with lower hydration kinetics (Figure 5.10). The ratio between the change in weight during cooking and the initial weight of pasta [(Wt-W0) / W0] is presented Table 5.6. There was a significant impact of the legume flour used to fortify pasta and of the treatment applied with no interaction between both of them. Whatever the treatment applied, cooked faba bean pasta presented a lower water absorption compared to split pea pasta (1.10 vs. 1.14, respectively). Whatever the legume pasta considered, VHT.LM pasta had a lower water absorption compared to LT pasta (1.02 vs. 1.16, respectively).
- 179 -
Chapitre 5. Pâtes alimentaires enrichies en légumineuse. Structure des pâtes et digestibilité de l’amidon
Table 5.6. Two way analysis of variance and subsequent LSD test on optimal cooking time, water absorption and in vitro rapidly available glucose (RAG) of split pea (SP) pasta and faba bean (FB) pasta cooked at their optimal cooking time (OCT) + 1min. Effect of legume Effect OCT +1min
df
A: Legume
1
0.29
B: Treatment
3
4)-beta-D-glucan rich fraction from barley on the physico-chemical properties and in vitro reducing sugars release of durum wheat pasta. International Journal of Food Science and Technology, 41(8), 910-918. Colonna, P., Barry, J. L., Cloarec, D., Bornet, F., Gouilloud, S., & Galmiche, J. P. (1990). Enzymic susceptibility of starch from pasta. Journal of Cereal Science, 11(1), 59-70. Colonna, P., Buleon, A., & Mercier, C. (1981). Pisum sativum and Vicia faba carbohydrates: structural studies of starches. Journal of Food Science, 46(1), 88-93. Colonna, P., Leloup, V., & Buleon, A. (1992). Limiting Factors of Starch Hydrolysis. European Journal of Clinical Nutrition, 46, S17-S32.
- 218 -
Références bibliographiques Colonna, P., & Mercier, C. (1979). [Leguminosae starches: aspect, composition, structure, and physico-chemical properties.]. Lebensmittel-Wissenschaft und -Technologie, 12(1), 1-12. Cubadda, R. E., Carcea, M., Marconi, E., & Trivisonno, M. C. (2007). Influence of gluten proteins and drying temperature on the cooking of durum wheat pasta. Cereal Chemistry, 84(1), 48-55. Cummings, J. H., & Englyst, H. N. (1995). Gastrointestinal Effects of Food Carbohydrate. American Journal of Clinical Nutrition, 61(4), S938-S945. Cunin, C., Handschin, S., Walther, P., & Escher, F. (1995). Structural changes of starch during cooking of durum wheat pasta. Lebensmittel-Wissenschaft und -Technologie, 28(3), 323-328. Cuq, B., Abecassis, J., & Guilbert, S. (2003). State diagrams to help describe wheat bread processing. International Journal of Food Science & Technology: 38 (7) 759-766, 38(7), 759-766. Cuq, J. L., & Leynaud-Rouaud, C. (1992). Les graines de légumineuses. In: H. Dupin, J. L. Cuq, C. Malewiak, C. Leynaud-Rouaud, & A. Berthier, Alimentation et Nutrition Humaine. Paris. Czuchajowska, Z., Otto, T., Paszczynska, B., & Byung-Kee, B. (1998). Composition, thermal behavior, and gel texture of prime and tailings starches from garbanzo beans and peas. Cereal Chemistry, 75(4), 466-472. Dalbon, G., Grivon, D., & Pagani, M. A. (1996). Pasta: Continuous manufacturing process. In: J. E. Kruger, R. R. Matsuo, & J. W. Dick, Pasta and Noodle Technology (pp. 13-58). St Paul, MN, USA: American Asscociation of Cereal Chemists. Davis, P. J. (1998). Protein modification by thermal processing. Allergy, 53(46), 102-105. de Almeida Costa, G. E., da Silva Queiroz-Monici, K., Pissini Machado Reis, S. M., & Costa de Oliveira, A. (2006). Chemical composition, dietary fibre and resistant starch contents of raw and cooked pea, common bean, chickpea and lentil legumes. Food Chemistry, 94(3), 327-330. de Stefanis, E., & Sgrulletta, D. (1990). Effects of high-temperature drying on technological properties of pasta. Journal of Cereal Science, 12(1), 97-104. De Zorzi, M., Curioni, A., Simonato, B., Giannattasio, M., & Pasini, G. (2007). Effect of pasta drying temperature on gastrointestinal digestibility and allergenicity of durum wheat proteins. Food Chemistry, 104(1), 353-363. Debbouz, A., Pitz, W. J., Moore, W. R., & Dappolonia, B. L. (1995). Effect of Bleaching on Durum-Wheat and Spaghetti Quality. Cereal Chemistry, 72(1), 128-131. Degidio, M. G., Mariani, B. M., Nardi, S., Novaro, P., & Cubadda, R. (1990). Chemical and technological variables and their relationships - a predictive equation for pasta cooking quality. Cereal Chemistry, 67(3), 275-281. Degidio, M. G., Mariani, B. M., & Novaro, P. (1993). Influence of Raw-Material Characteristics and Drying Technologies on Pasta Cooking Quality - a Review of Our Results. Industrie Alimentari, 29-32. del Nobile, M. A., Baiano, A., Conte, A., & Mocci, G. (2005). Influence of protein content on spaghetti cooking quality. Journal of Cereal Science, 41(3), 347-356. del Nobile, M. A., & Massera, M. (2000). Modeling of water sorption kinetics in spaghetti during overcooking. Cereal Chemistry, 77(5), 615-619. Devaux, M. F., Robert, P., Melcion, J. P., & deMonredon, F. L. (1997). Particle size analysis of bulk powders using mathematical morphology. Powder Technology, 90(2), 141-147. Dexter, J. E., Dronzek, B. L., & Matsuo, R. R. (1978). Scanning electron microscopy of cooked spaghetti. Cereal Chemistry, 55(1), 23-30. Dexter, J. E., & Matsuo, R. R. (1977). Changes in semolina proteins during spaghetti processing. Cereal Chemistry: 54 (4) 882-894, 54(4), 882-894.
- 219 -
Références bibliographiques Dexter, J. E., & Matsuo, R. R. (1979). Effect of starch on pasta dough rheology and spaghetti cooking quality. Cereal Chemistry, 56(3), 190-195. Dexter, J. E., Tkachuk, R., & Matsuo, R. R. (1984). Amino acid composition of spaghetti: effect of drying conditions on total and available lysine. Journal of Food Science, 49(1), 225-228. Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A., & Smith, F. (1956). Colorimetric method for determination of sugars and related substances. Analytical Chemistry, 28, 350-356. Duranti, M. (2006). Grain legume proteins and nutraceutical properties. Fitoterapia, 77(2), 67-82. Edwards, N. M., Biliaderis, C. G., & Dexter, J. E. (1995). Textural characteristics of wholewheat pasta and pasta containing non-starch polysaccharides. . Journal of Food Science, 60(6), 1321-1324. El-Fiel, H. E. A., El-Tinay, A. H., & Elsheikh, E. A. E. (2002). Effect of nutritional status of faba bean (Vicia faba L.) on protein solubility profiles. Food Chemistry, 76(2), 219-223. El-Khayat, G., Samaan, J., & Brennan, C. S. (2003). Evaluation of vitreous and starchy Syrian durum (Triticum durum) wheat grains: the effect of amylose content on starch characteristics and flour pasting properties. Starch/Staerke, 55(8), 358-365. Englyst, H. N., Kingman, S. M., & Cummings, J. H. (1992). Classification and Measurement of Nutritionally Important Starch Fractions. European Journal of Clinical Nutrition, 46, S33-S50. Englyst, K. N., Englyst, H. N., Hudson, G. J., Cole, T. J., & Cummings, J. H. (1999). Rapidly available glucose in foods: an in vitro measurement that reflects the glycemic response. American Journal of Clinical Nutrition, 69(3), 448-454. Epstein, J., Morris, C. F., & Huber, K. C. (2002). Instrumental texture of white salted noodles prepared from recombinant inbred lines of wheat differing in the three granule bound starch synthase (Waxy) genes. Journal of Cereal Science, 35(1), 51-63. FAO/WHO/UNU (1985). Energy and protein requirements. WHO technical report series No. 724. Geneva: World Health Organization. Fardet, A., Abecassis, J., Hoebler, C., Baldwin, P. M., Buleon, A., Berot, S., & Barry, J. L. (1999). Influence of technological modifications of the protein network from pasta on in vitro starch degradation. Journal of Cereal Science, 30(2), 133-145. Fardet, A., Baldwin, P. M., Bertrand, D., Bouchet, B., Gallant, D. J., & Barry, J. L. (1998a). Textural images analysis of pasta protein networks to determine influence of technological processes. Cereal Chemistry, 75(5), 699704. Fardet, A., Hoebler, C., Baldwin, P. M., Bouchet, B., Gallant, D. J., & Barry, J. L. (1998b). Involvement of the protein network in the in vitro degradation of starch from spaghetti and lasagne: a microscopic and enzymic study. Journal of Cereal Science, 27(2), 133-145. Favier, J., Samson, M. F., Aubled, C., Morel, M. H., & Abecassis, J. (1996). Thermal kinetics of dough proteins insolubilisation determined by size exclusion high performance liquid chromatography. Sciences des Aliments: 16 (6) 573-591, 16(6), 573-591. Feillet, P. (2000). Le grain de blé: composition et utilisation. INRA edition. Feillet, P., & Dexter, J. E. (1996). Quality requirements of durum wheat for semolina milling and pasta production. In: J. E. Kruger, R. R. Matsuo, & J. W. Dick, Pasta and Noodle Technology (pp. 95-131). St Paul, MN, USA: American Association of Cereal Chemists. Fosterpowell, K., & Miller, J. B. (1995). International Tables of Glycemic Index. American Journal of Clinical Nutrition, 62(4), S871-S890.
- 220 -
Références bibliographiques Frias, J., Kovacs, E., Sotomayor, C., Hedley, C., & Vidal-Valverde, C. (1997). Processing peas for producing macaroni. Zeitschrift fuer Lebensmittel-Untersuchung und-Forschung A/Food Research and Technology, 204(1), 66-71. Goni, I., & Valentin-Gamazo, C. (2003). Chickpea flour ingredient slows glycemic response to pasta in healthy volunteers. Food Chemistry, 81(4), 511-515. Granfeldt, Y., & Bjorck, I. (1991a). Glycemic response to starch in pasta: a study of mechanisms of limited enzyme availability. Journal of Cereal Science, 14, 47-61. Granfeldt, Y., Bjorck, I., & Hagander, B. (1991b). On the Importance of Processing Conditions, Product Thickness and Egg Addition for the Glycemic and Hormonal Responses to Pasta - a Comparison with Bread Made from Pasta Ingredients. European Journal of Clinical Nutrition, 45(10), 489-499. Grant, L. A., Dick, J. W., & Shelton, D. R. (1993). Effects of drying temperature, starch damage sprouting, and additives on spaghetti quality characteristics. Cereal Chemistry, 70(6), 676-684. Grzybowski, R. A., & Donnelly, B. J. (1977). Starch gelatinization in cooked spaghetti. Journal of Food Science, 42(5), 1304-1305. Gueguen, J., & Barbot, J. (1988). Quantitative and qualitative variability of pea (Pisum sativum L.) protein composition. Journal of the Science of Food and Agriculture, 42(3), 209-224. Guillemin, F. (2003). Morphologie des cellules de la racine de betterave sucrière et distribution des épitopes pectiques. (p. 176). Nantes: Université de Nantes. Guillon, F., & Champ, M. M. J. (2002). Carbohydrate fractions of legumes: uses in human nutrition and potential for health. British Journal of Nutrition, 88, S293-S306. Guler, S., Koksel, H., & Ng, P. K. W. (2002). Effects of industrial pasta drying temperatures on starch properties and pasta quality. Food Research International, 35(5), 421-427. Hansen, L., & Johnston, P. (1976). Heat-moisture effects on wheat flour.2. An evalutaion study of heat-processing effects on flour proteins by digestive enzymes-pepsin, trypsin, and trypsin-carboxypeptidase-B. Cereal Chemistry, 53(5), 656-670. Hefle, S. L., Lemanske, R. F., & Bush, R. K. (1994). Adverse Reaction to Lupine-Fortified Pasta. Journal of Allergy and Clinical Immunology, 94(2), 167-172. Heneen, W. K., & Brismar, K. (2003). Structure of cooked spaghetti of durum and bread wheats. Starch-Starke, 55(12), 546-557. Hoebler, C., Devaux, M. F., Karinthi, A., Belleville, C., & Barry, J. L. (2000). Particle size of solid food after human mastication and in vitro simulation of oral breakdown. International Journal of Food Sciences and Nutrition, 51(5), 353-366. Holm, J., Bjoerck, I., & Eliasson, A. C. (1988a). Effects of thermal processing of wheat on starch. I. Physicochemical and functional properties. Journal of Cereal Science: 8 (3) 249-260, 8(3), 249-260. Holm, J., & Bjorck, I. (1988b). Effects of Thermal-Processing of Wheat on Starch .2. Enzymic Availability. Journal of Cereal Science, 8(3), 261-268. Holm, J., Bjorck, I., Ostrowska, S., Eliasson, A. C., Asp, N. G., Larsson, K., & Lundquist, I. (1983). Digestibility of Amylose-Lipid Complexes Invitro and Invivo. Starke, 35(9), 294-297. Horigane, A. K., Naito, S., Kurimoto, M., Irie, K., Yamada, M., Motoi, H., & Yoshida, M. (2006). Moisture distribution and diffusion in cooked spaghetti studied by NMR imaging and diffusion model. Cereal Chemistry, 83(3), 235-242.
- 221 -
Références bibliographiques Hospers, J. J., van Amelsvoort, J. M. M., & Weststrate, J. A. (1994). Amylose-to-amylopectin ratio in pastas affects postprandial glucose and insulin responses and satiety in males. Journal of Food Science, 59(5), 1144-1149. Huang, J., Schols, H. A., van Soest, J. G., Jin, Z., Sulmann, E., & Voragen, A. G. J. (2007). Physicochemical properties and amylopectin chain profiles. Food Chemistry, 101, 1355-1362. Icard-Vernière, C. (1999). De la semoule de blé dur aux pâtes alimentaires fraiches: évènements physiques et biochimiques. (p. 183). Montpellier: Université Montpellier II. Ingelbrecht, J. A., Moers, K., Abecassis, J., Rouau, X., & Delcour, J. A. (2001). Influence of arabinoxylans and endoxylanases on pasta processing and quality. Production of high-quality pasta with increased levels of soluble fiber. Cereal Chemistry, 78(6), 721-729. Jacobs, H., Eerlingen, R. C., Rouseu, N., Colonna, P., & Delcour, J. A. (1998). Acid hydrolysis of native and annealed wheat, potato and pea starches - DSC melting features and chain lenght distributions of lintnerised starches. Carbohydrate Research, 308(3-4), 359-371. Jenkins, D. J. A., Kendall, C. W. C., Augustin, L. S. A., Franceschi, S., Hamidi, M., Marchie, A., Jenkins, A. L., & Axelsen, M. (2002). Glycemic index: overview of implications in health and disease. American Journal of Clinical Nutrition, 76(1), 266S-273S. Jenkins, D. J. A., Taylor, R. H., & Wolever, T. M. S. (1982). The Diabetic Diet, Dietary Carbohydrate and Differences in Digestibility. Diabetologia, 23(6), 477-484. Jenkins, D. J. A., Wolever, T. M. S., Jenkins, A. L., Lee, R., Wong, G. S., & Josse, R. G. (1983). Glycemic response to wheat products: reduced response to pasta but no effect of fiber. Diabetes care, 6, 155-159. Jenkins, D. J. A., Wolever, T. M. S., Taylor, R. H., Barker, H., Fielden, H., Baldwin, J. M., Bowling, A. C., Newman, H. C., Jenkins, A. L., & Goff, D. V. (1981). Glycemic index of foods: a physiological basis for carbohydrate exchange. The American Journal of Clinical Nutrition, 34, 362-366. Jood, S., Bishnoi, S., & Sharma, A. (1998). Chemical analysis and physico-chemical properties of chickpea and lentil cultivars. Nahrung, 42(2), 71-74. Kies, C., & Fox, H. M. (1970). Determination of the first-limiting amino acid of wheat and triticale grain for humans. Cereal Chemistry, 47, 615 - 625. Kim, E. H. J., Petrie, J. R., Motoi, L., Morgenstern, M. P., Sutton, K. H., Suman, M., & Simmons, L. D. (2008). Effect of structural and physicochemical characteristics of the protein matrix in pasta on in vitro starch digestibility. Food Biophysics: 3 (2) 229-234, 3(2), 229-234. Klepacka, M., Porzucek, H., & Kluczynska, M. (1997). Effect of heat treatment on chemically modified proteins of legume seeds. Food Chemistry, 58(3), 219-222. Kokini, J. L., Cocero, A. M., Madeka, H., & de Graaf, E. (1994). The development of state diagrams for cereal proteins. Trends in Food Science & Technology: 5 (9) 281-288, 5(9), 281-288. Koyoro, H., & Powers, J. R. (1987). Functional properties of pea globulin fractions. Cereal Chemistry, 64(2), 97101. Kruger, J. E., R.B., M., & J.W., D. (1996). Pasta and Noodle Technology. St Paul, Minnesota: American Association of Cereal Chemists. Lamacchia, C., di Luccia, A., Baiano, A., Gambacorta, G., la Gatta, B., Pati, S., & la Notte, E. (2007). Changes in pasta proteins induced by drying cycles and their relationship to cooking behaviour. Journal of Cereal Science, 46(1), 58-63. Lampart-Szczapa, E., Obuchowski, W., Czaczyk, K., Pastuszewska, B., & Buraczewska, L. (1997). Effect of lupine flour on the quality and oliosaccharides of pasta and crisps. Nahrung: 41 (4) 219-223, 41(4), 219-223.
- 222 -
Références bibliographiques Landillon, V., Cassan, D., Morel, M. H., & Cuq, B. (2008). Flowability, cohesive, and granulation properties of wheat powders. Journal of Food Engineering: 86 (2) 178-193, 86(2), 178-193. Leterme, P., Monmart, T., & Baudart, E. (1990). Amino acid composition of pea (Pisum sativum) proteins and protein profile of pea flour. Journal of the Science of Food and Agriculture, 53(1), 107-110. Li, M., & Lee, A. C. (2000). Effect of extrusion temperature on the solubility and molecular weight of lentil bean flour proteins containing low cysteine residues. Journal of Agricultural and Food Chemistry, 48(3), 880-884. Lintas, C., & d' Appolonia, B. L. (1973). Effect of spaghetti processing on semolina carbohydrates. Cereal Chemistry, 50(5), 563-570. Lorenz, K. (1979). The starch of the Fababean (Vicia faba). Comparison with wheat and corn starch. Staerke, 31(6), 181-184. Lucisano, M., Pagani, M. A., Mariotti, M., & Locatelli, D. P. (2008). Influence of die material on pasta characteristics. Food Research International: 41 (6) 646-652, 41(6), 646-652. Manthey, F. A., & Hall, C. A., II (2007). Effect of processing and cooking on the content of minerals and protein in pasta containing buckwheat bran flour. Journal of the Science of Food and Agriculture, 87(11), 2026-2033. Manthey, F. A., & Schorno, A. L. (2002). Physical and cooking quality of spaghetti made from whole wheat durum. Cereal Chemistry, 79(4), 504-510. Manthey, F. A., Yalla, S. R., Dick, T. J., & Badaruddin, M. (2004). Extrusion properties and cooking quality of spaghetti containing buckwheat bran flour. Cereal Chemistry, 81(2), 232-236. Marshall, S., & Wasik, R. (1974). Gelatinization of starch during cooking of spaghetti. Cereal Chemistry, 51, 146147. Matsuo, R. R., Dexter, J. E., & Dronzek, B. L. (1978). Scanning electron microscopy study of spaghetti processing. Cereal Chemistry, 55(5), 744-753. Micard, V., Morel, M. H., Bonicel, J., & Guilbert, S. (2001). Thermal properties of raw and processed wheat gluten in relation with protein aggregation. Polymer, 42(2), 477-485. Miller, J. B., Pang, E., & Bramall, L. (1992). Rice - a High or Low Glycemic Index Food. American Journal of Clinical Nutrition, 56(6), 1034-1036. Mohamed, A. A., & Rayasduarte, P. (1995). Composition of Lupinus-Albus. Cereal Chemistry, 72(6), 643-647. Mohamed, S. (1990). Factors Affecting Extrusion Characteristics of Expanded Starch-Based Products. Journal of Food Processing and Preservation, 14(6), 437-452. Monge, L., Cortassa, G., Fiocchi, F., Mussino, G., & Carta, Q. (1990). Glyco-insulinaemic response, digestion and intestinal absorption of the starch contained in two types of spaghetti. . Diabetes, Nutrition and Metabolism, 3, 239246. Morel, M. H., Dehlon, P., Autran, J. C., Leygue, J. P., & Bar-L'Helgouac'h, C. (2000). Effects of temperature, sonication time, and power settings on size distribution and extractability of total wheat flour proteins as determined by size-exclusion high-performance liquid chromatography. Cereal Chemistry, 77(5), 685-691. Mujoo, R., & Ali, S. Z. (1998). Susceptibility of starch to in vitro enzyme hydrolysis in rice, rice flakes and intermediary products. Food Science and Technology-Lebensmittel-Wissenschaft & Technologie, 31(2), 114-121. Nielsen, M. A., Sumner, A. K., & Whalley, L. L. (1980). Fortification of pasta with pea flour and air-classified pea protein concentrate. Cereal Chemistry, 57(3), 203-206. Novaro, P., d' Egidio, M. G., Mariani, B. M., & Nardi, S. (1993). Combined effect of protein content and hightemperature drying systems on pasta cooking quality. Cereal Chemistry, 70(6), 716-719.
- 223 -
Références bibliographiques Nur Herken, E., Ibanog lu, S., Oener, M. D., & Ibanog Iu, E. (2006). The in vitro protein digestibility, microbiological quality and gelatinization behaviour of macaroni as affected by cowpea flour addition. Food Chemistry, 98(4), 664-669. O'Kane, F. E., Happe, R. P., Vereijken, J. M., Gruppen, H., & Van Boekel, M. (2004). Heat-induced gelation of pea legumin: Comparison with soybean glycinin. Journal of Agricultural and Food Chemistry, 52(16), 5071-5078. Oliver, J. R., Blakeney, A. B., & Allen, H. M. (1993). The color of flour streams as related to ash and pigment contents. Journal of Cereal Science, 17(2), 169-182. Olivera, D. F., & Salvadori, V. O. (2006). Textural characterisation of lasagna made from organic whole wheat. International Journal of Food Science and Technology, 41, 63-69. Opperman, A. M., Venter, C. S., Oosthuizen, W., Thompson, R. L., & Vorster, H. H. (2004). Meta-analysis of the health effects of using the glycaemic index in meal-planning. British Journal of Nutrition, 92(3), 367-381. Osborne, T. B. (1907). The proteins of the wheat kernel. Osorio-Diaz, P., Agama-Acevedo, E., Mendoza-Vinalay, M., Tovar, J., & Bello-Perez, L. A. (2008). Pasta added with chickpea flour: Chemical composition, in vitro starch digestibility and predicted glycemic index. Ciencia Y Tecnologia Alimentaria, 6(1), 6-12. Pagani, M. A., Gallant, D., Bouchet, B., & Resmini, P. (1986). Ultrastructure of cooked spaghetti. Food Microstructure, 5, 111-129. Pagani, M. A., Resmini, P., & Dalbon, G. (1989). Influence of the extrusion process on characteristics and structure of pasta. Food Microstructure, 8(2), 173-182. Parker, R., & Ring, S. G. (2001). Aspects of the physical chemistry of starch. Journal of Cereal Science, 34(1), 117. Pasini, G., Simonato, B., Giannattasio, M., Peruffo, A. D. B., & Curioni, A. (2001). Modifications of wheat flour proteins during in vitro digestion of bread dough, crumb, and crust: An electrophoretic and immunological study. Journal of Agricultural and Food Chemistry, 49(5), 2254-2261. Pasqualini, S., Lluch, C., & Antonielli, M. (1991). Seed storage proteins in several genetic lines of Vicia faba. Plant Physiology and Biochemistry, 29(5), 507-515. Perez-Hidalgo, M. A., Guerra-Hernandez, E., & Garcia-Villanova, B. (1997). Dietary Fiber in Three Raw Legumes and Processing Effect on Chick Peas by an Enzymatic-Gravimetric Method. Journal of Food Composition and Analysis, 10(1), 66-72. Petitot, M., Abecassis, J., & Micard, V. (2009a). Structuring of pasta components during processing: Impact on starch and protein digestibility and allergenicity. Trends in Food Science & Technology. Petitot, M., Barron, C., Morel, M. H., & Micard, V. (2009b). Impact of legume addition on pasta structure: consequences on in vitro starch digestibility. Submitted to Food Biophysics Petitot, M., Brossard, C., Barron, C., Larré, C., Morel, M. H., & Micard, V. (2009c). Modification of pasta structure induced by high drying temperatures. Effects on the in vitro digestibility of protein and starch fractions and the potential allergenicity of protein hydrolysates. Food Chemistry, 116(2), 401-412. Petitot, M., & Micard, V. (2009d). Fortification of pasta with legume flour: processing ease, cooking quality, structure and nutritional properties Petitot, M., & Micard, V. (2009e). Fortification of pasta with split pea and faba bean flours : Pasta processing and quality evaluation. Food Research International. Petitot, M., & Micard, V. (2009f). Legume fortified pasta. Impact of drying and precooking treatments on pasta structure and inherent starch digestibility. Journal of Cereal Science.
- 224 -
Références bibliographiques Pollini, C. M. (1996). THT technology in the modern industrial pasta making process. In: J. E. Kruger, R. R. Matsuo, & J. W. Dick, Pasta Noodle and Technology (pp. 59-74). St Paul, MN: American Association of Cereal Chemists. Rahman, M. S. (2008). Dehydration and microstructure. In: D.-W. Sun, Advances in food dehydration (pp. 97-123). St-Foy: CRC Press. Rasanen, L., Lehto, M., Turjanmaa, K., Savolainen, J., & Reunala, T. (1994). Allergy to Ingested Cereals in Atopic Children. Allergy, 49(10), 871-876. Rayas-Duarte, P., Mock, C. M., & Satterlee, L. D. (1996). Quality of spaghetti containing buckwheat, amaranth, and lupin flours. Cereal Chemistry, 73(3), 381-387. Redl, A., Morel, M. H., Bonicel, J., Vergnes, B., & Guilbert, S. (1999). Extrusion of wheat gluten plasticized with glycerol: Influence of process conditions on flow behavior, rheological properties, and molecular size distribution. Cereal Chemistry, 76(3), 361-370. Rehman, Z., & Shah, W. H. (2005). Thermal heat processing effects on antinutrients, protein and starch digestibility of food legumes. Food Chemistry, 91(2), 327-331. Resmini, P., & Pellegrino, L. (1994). Occurrence of protein-bound lysylpyrrolaldehyde in dried pasta. Cereal Chemistry: 71 (3) 254-262, 71(3), 254-262. Rouille, J., Della Valle, G., Devaux, M. F., Marion, D., & Dubreil, L. (2005). French bread loaf volume variations and digital image analysis of crumb grain changes induced by the minor components of wheat flour. Cereal Chemistry, 82(1), 20-27. Sabanis, D., Makri, E., & Doxastakis, G. (2006). Effect of durum flour enrichment with chickpea flour on the characteristics of dough and lasagne. Journal of the Science of Food and Agriculture, 86(12), 1938-1944. Sadeghi, M. A., & Bhagya, S. (2008). Quality characterization of pasta enriched with mustard protein isolate. Journal of Food Science, 73(5), S229-S237. Sandin, D. I., San Ireneo, M. M., Lizana, F. M., Fernandez-Caldas, E., Lebrero, E. A., & Borrego, T. L. (1999). Specific IgE determinations to crude and boiled lentil (Lens culinaris) extracts in lentil-sensitive children and controls. Allergy, 54(11), 1209-1214. Schofield, J. D. (1986). Flour proteins: Structure and functionality in baked products. In: J. M. V. Blanshard, P. J. Frazier, & T. Galliard, Chemistry and Physics of Baking (pp. 14-29). UK: The Royal Society of Chemistry. Sell, M., Steinhart, H., & Paschke, A. (2005). Influence of maturation on the alteration of allergenicity of green pea (Pisum sativum L.). Journal of Agricultural and Food Chemistry, 53(5), 1717-1722. Simonato, B., Pasini, G., De Zorzi, M., Vegro, M., & Curioni, A. (2004). Potential allergens in durum wheat semolina and pasta: Fate during cooking and digestion. Italian Journal of Food Science, 16(2), 151-163. Simonato, B., Pasini, G., Giannattasio, M., Peruffo, A. D. B., De Lazzari, F., & Curioni, A. (2001). Food allergy to wheat products: The effect of bread baking and in vitro digestion on wheat allergenic proteins. A study with bread dough, crumb, and crust. Journal of Agricultural and Food Chemistry, 49(11), 5668-5673. Singh, H., & MacRitchie, F. (2004). Changes in proteins induced by heating gluten dispersions at high temperature. Journal of Cereal Science, 39(2), 297-301. Singh, N., Singh, J., Kaur, L., Sodhi, N. S., & Gill, B. S. (2003). Morphological, thermal and rheological properties of starches from different botanical sources. Food Chemistry, 81(2), 219-231. Singh, N. K., Donovan, G. R., Batey, I. L., & MacRitchie, F. (1990). Use of sonication and size-exclusion highperformance liquid chromatography in the study of wheat flour proteins. I. Dissolution of total proteins in the absence of reducing agents. Cereal Chemistry: 67 (2) 150-161, 67(2), 150-161.
- 225 -
Références bibliographiques Souci, S. W., Fachmann, W., & Kraut, H. (1989). Food composition and nutrition tables. Stuttgart: Medpharm GmbH Scientific Publishers. Sujak, A., Kotlarz, A., & Strobel, W. (2006). Compositional and nutritional evaluation of several lupin seeds. Food Chemistry, 98(4), 711-719. Svihus, B., Uhlen, A. K., & Harstad, O. M. (2005). Effect of starch granule structure, associated components and processing on nutritive value of cereal starch: A review. Animal Feed Science and Technology, 122(3-4), 303-320. Tharanathan, R. N., & Mahadevamma, S. (2003). Grain legumes - a boon to human nutrition. Trends in Food Science & Technology: 14 (12) 507-518, 14(12), 507-518. Torres, A., Frias, J., Granito, M., Guerra, M., & Vidal-Valverde, C. (2007a). Chemical, biological and sensory evaluation of pasta products supplemented with alpha-galactoside-free lupin flours. Journal of the Science of Food and Agriculture, 87(1), 74-81. Torres, A., Frias, J., Granito, M., & Vidal-Valverde, C. (2007b). Germinated Cajanus cajan seeds as ingredients in pasta products: Chemical, biological and sensory evaluation. Food Chemistry, 101(1), 202-211. Tudorica, C. M., Kuri, V., & Brennan, C. S. (2002). Nutritional and physicochemical characteristics of dietary fiber enriched pasta. Journal of Agricultural and Food Chemistry: 50 (2) 347-356, 50(2), 347-356. Vansteelandt, J., & Delcour, J. A. (1998). Physical behavior of durum wheat starch (Triticum durum) during industrial pasta processing. Journal of Agricultural and Food Chemistry, 46(7), 2499-2503. Vansteelandt, J., & Delcour, J. A. (1999). Characterisation of starch from durum wheat (Triticum durum). Starch/Staerke, 51(2/3), 73-80. Varjonen, E., Vainio, E., & Kalimo, K. (1997). Life-threatening, recurrent anaphylaxis caused by allergy to gliadin and exercise. Clinical and Experimental Allergy, 27(2), 162-166. Varjonen, E., Vainio, E., & Kalimo, K. (2000). Antigliadin IgE - indicator of wheat allergy in atopic dermatitis. Allergy, 55(4), 386-391. Visschers, R. W., & de Jongh, H. H. J. (2005). Disulphide bond formation in food protein aggregation and gelation. Biotechnology Advances, 23(1), 75-80. Wal, J. M. (2003). Thermal processing and allergenicity of foods. Allergy, 58(8), 727-729. Weegels, P. L., & Hamer, R. J. (1998). Temperature induced changes of wheat products. . In: R. J. Hamer, & R. C. Hoseney, Interactions: the key to cereal quality (pp. 95-130). Saint Paul, MN: American Association of Cereal Chemists. Weegels, P. L., Verhoek, J. A., Degroot, A. M. G., & Hamer, R. J. (1994). Effects on Gluten of Heating at Different Moisture Contents .1. Changes in Functional-Properties. Journal of Cereal Science, 19(1), 31-38. Wolever, T. M. S., Jenkins, D. J. A., Kalmusky, J., Giordano, C., Giudici, S., Jenkins, A. L., Thompson, L. U., Wong, G. S., & Josse, R. G. (1986). Glycemic Response to Pasta: Effect of Surface area, Degree of Cooking, and Protein Enrichment. Diabetes care, 9(4), 401-404. Wood, J. A. (2009). Texture, processing and organoleptic properties of chickpea-fortified spaghetti with insights to the underlying mechanisms of traditional durum pasta quality. Journal of Cereal Science, 49, 128-133. Yue, P., Rayas-Duarte, P., & Elias, E. (1999). Effect of drying temperature on physicochemical properties of starch isolated from pasta. Cereal Chemistry, 76(4), 541-547. Zardetto, S., & dalla Rosa, M. (2005). Study of the effect of lamination process on pasta by physical chemical determination and near infrared spectroscopy analysis. Journal of Food Engineering, 74(3), 402-409.
- 226 -
Références bibliographiques 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 flours. Journal of Food Science, 70(6), S371-S376. Zheng, G. H., Fasina, O., Sosulski, F. W., & Tyler, R. T. (1998). Nitrogen solubility of cereals and legumes subjected to micronization. Journal of Agricultural and Food Chemistry, 46(10), 4150-4157. Zweifel, C. (2001). Influence of high-temperature drying on the structural and textural properties of durum wheat pasta. Zurich: Swiss Federal Institute of Technology. Zweifel, C., Conde-Petit, B., & Escher, F. (2000). Thermal modifications of starch during high-temperature drying of pasta. Cereal Chemistry, 77(5), 645-651. Zweifel, C., Handschin, S., Escher, F., & Conde-Petit, B. (2003). Influence of high-temperature drying on structural and textural properties of durum wheat pasta. Cereal Chemistry, 80(2), 159-167.
- 227 -
- 228 -
Travaux relatifs à cette étude
- 229 -
- 230 -
Travaux relatifs à cette étude
Publications dans des revues internationales à comité de lecture Petitot, M., Abecassis, J., & Micard, V. (2009a). Structuring of pasta components during processing: Impact on starch and protein digestibility and allergenicity. Revue acceptée pour publication dans Trends in Food Science & Technology. Petitot, M., Barron, C., Morel, M. H., & Micard, V. (2009b). Structure of pasta fortified with split pea and faba bean flours: impact on starch digestibility. Soumise à Food Chemistry. Petitot, M. Brossard, C. Barron, C., Larré, C. Morel, M.H., and Micard, V. (2009c). Modification of pasta structure induced by high drying temperatures. Effects on the in vitro digestibility of protein and starch fractions and the potential allergenicity of protein hydrolysates. Food Chemistry, 116(2): 401-412 Petitot, M and Micard, V. (2009d). Fortification of pasta with legume flour: processing ease, cooking quality, structure and nutritional properties. Revue en cours de préparation. Petitot, M., & Micard, V. (2009e). Fortification of pasta with split pea and faba bean flours : Pasta processing and quality evaluation. Acceptée pour publication dans Food Research International. Petitot, M., Micard, V. (2009f).Legume fortified pasta. Impact of drying and precooking treatments on pasta structure and starch digestibilit. A soumettre à Journal of Cereal Science
Communications orales à des congrès scientifiques Petitot, M., Brossard, C., Barron, C., Larré, C., Morel, M.H., Micard, V. 2008. Effect of high drying temperatures on durum wheat pasta structure. Impact on protein and starch digestibility. Congrès 13th ICC Cereal and Bread Congress, Cerworld 21st, 15-18 Juin 2008. Madrid. Espagne. Petitot, M., Cassan, D., Morel, M.H., Cuq, B. Micard, V. PASTALEG : design of pasta made from wheat and legume. Impact of the food matrix structure on its nutritional properties. 6th European Young Cereal Scientists and Technologists Workshop, April 31, 2007, Montpellier, France. Abecassis, J, Petitot, M., Micard, V. 2007. Effect of Milling and processing on sensorial and nutritional properties of pasta products. 10th Federation of European Nutrition Societies (FENS) conference, 10-13 July 2007. Paris. France.
- 231 -
Communications affichées à des congrès scientifiques Petitot, M. Brossard, C., Barron, C., Larré, C. Morel, M.H., Micard, V. 2008. Changes in pasta protein and starch structures induced by drying and their relationship to protein and starch digestibilities. From Seed to patsa, 30 June – 3 July 2008, Bologna, Italy. Petitot, M., Morel, M.H., Brossard, C., Champ, M., Crenon, I., Minier, C., Micard, V. 2007. Association of durum wheat and legume into food matrix. Impact on its nutritional value. Federation of European Nutrition Societies (FENS), 10-13 July 2007. Paris. France.
- 232 -
Pâtes alimentaires enrichies en légumineuse. Structuration des constituants au cours du procédé : impact sur la qualité culinaire et les propriétés nutritionnelles des pâtes RESUME : Afin d’approfondir les connaissances sur le lien structure/nutrition, des pâtes de structures constatées ont été obtenues en modifiant la formulation et/ou le procédé de fabrication des pâtes. Les propriétés structurales des pâtes ainsi obtenues ont été étudiées à différentes échelles, puis reliées à la digestibilité in vitro de l’amidon. L’introduction de 35% de farine de fève ou de pois cassé dans les pâtes au blé dur engendre des changements structuraux mineurs des pâtes sans toutefois affecter la digestibilité in vitro de l’amidon. Des phénomènes antagonistes favorisant la digestibilité de l’amidon ou au contraire l’inhibant pourraient expliquer ce phénomène. L’application de différents traitements technologiques sur les pâtes enrichies en légumineuse permet d’obtenir les modifications structurales les plus importantes, engendrant de ce fait une modification marquée de la digestibilité in vitro de l’amidon. En accord avec les hypothèses avancées dans la littérature, ces travaux de thèse ont démontré que la faible digestibilité de l’amidon dans les pâtes pouvait s’expliquer par la compacité de sa structure (une augmentation de la porosité entraîne bien une digestibilité accrue) et la présence d’un réseau protéique protégeant les granules d’amidon de l’attaque enzymatique. D’après nos résultats, le rôle protecteur du réseau serait davantage lié à la présence de nombreuses liaisons covalentes (ponts disulfures et autres liaisons covalentes fortes) au niveau supramoléculaire plutôt qu’à la distribution spatiale des protéines au niveau microscopique. La présence d’un réseau protéique très réticulé serait plus résistant à l’attaque protéasique ce qui pourrait ralentir l’accès des amylases à l’amidon. La faible digestibilité de l’amidon dans les pâtes peut donc être réduite davantage par un enrichissement avec 35% de farine de légumineuse combiné à l’utilisation de très hautes températures de séchage. Cependant, l’impact de tels traitements sur la digestibilité protéique, la biodisponibilité de la lysine et l’allergénicité restent à déterminer. Mots clefs : pâtes au blé dur, légumineuse, structure, réseau protéique, digestibilité de l’amidon.
Legume fortified pasta. Structuring of pasta components during processing: impact on cooking quality and nutritional properties of pasta SUMMARY: In order to deepen the knowledge about structure and nutrition, different structured pasta were obtained by changing the formulation and/or pasta processing. Structural properties of pasta obtained were studied at different scales, and then connected to the in vitro digestibility of starch. Inclusion of 35% of faba bean or split pea flour in pasta induces minor structural changes without affecting the in vitro digestibility of starch. The compensation of antagonistic phenomena favoring or inhibiting starch digestibility could explain this result. The application of different technological treatments on legume pasta induces the most important structural changes, leading to important changes in the in vitro digestibility of starch. In accordance with hypothesis previously suggested by other authors, the low glycemic index of pasta could be due to the compactness of pasta structure (an increase in pasta porosity induces an increase in starch digestibility) and the presence of a protein network protecting starch from enzymatic attack. According to our results, the presence of a large amount of covalent links (disulfide bonds and other covalent bonds) at a macromolecular level would be more important than the spatial distribution of the protein network at a microscopic scale to control the digestibility of starch. A highly aggregated protein network would be more resistant to protease hydrolysis, which could delay the amylase hydrolysis of starch. The low digestibility of starch can thus be further reduced by the fortification of pasta with 35% legume flour combined to the use of very high drying temperatures. However, despite this interesting nutritional result, the impact of such treatments on the digestibility of proteins, biodisponibility of lysine and allergenicity should be determined in the future. Keywords: durum wheat pasta, legume, structure, protein network, starch digestibility
