Fortification of meat analogues with iron-loaded double emulsions: improving oxidative stability and development of an injection method

Transcription

1 Fortification of meat analogues with iron-loaded double emulsions: improving oxidative stability and development of an injection method Master Thesis 2018 Name: Study program: Peijun Peng Master Food Technology Reg. number: Starting date: September 2017 March 2018 Supervisors: Examiner: Patricia Duque Estrada, Claire Berton-Carabin Atze Jan van der Goot 1

2 Abstract Firstly, the physiochemical properties of double emulsions (W1/O/W2) prepared with olive oil and sunflower oil were studied and compared, in order to choose an oil which gives higher stability to W1/O/W2 to encapsulate iron. The physical stability was measured in terms of droplet size distribution, morphology, encapsulation efficiency (EE%) and creaming. The oxidative stability was determined by peroxide value, conjugated dienes value and para- Anisidine value. Both physical and chemical stability were measured in fresh samples after 1 day, 7 days and 14 days of storage. Secondly, the effect of α-tocopherol (100 ppm, 200 ppm, 300 ppm) to reduce lipid oxidation was investigated. Lastly, double emulsions with olive oil were injected into meat analogues using a simple syringe (0.5mm*16mm). Then, carbonyl content was measured as a chemical stability parameter in fresh samples after 1 day and 7 days of storage. Olive oil W1/O/W2 and sunflower oil W1/O/W2 had similar physical stability regarding droplet size distribution, morphology, EE% and creaming during 14 days of storage. High encapsulation efficiency was observed in both types of emulsions during storage. Moreover, olive oil W1/O/W2 had significantly higher oxidative stability than sunflower oil W1/O/W2, particularly in the formation of secondary lipid oxidation compounds. The lipid oxidation of olive oil W1/O/W2 can be effectively retarded by a- tocopherol at a concentration of 300 ppm. As for the injection of double emulsions to meat analogues, the leaking occurred when the simple syringe was applied. Carbonyl content did not differ significantly between samples with injected double emulsions and without emulsions, which indicated that the simple syringe might be inefficient to inject the W1/O/W2 into the meat analogues, or the amount of emulsions did not influence the oxidation in meat analogues. 2

3 Table of Contents Abstract Introduction Objective Theoretical background... 6 Iron... 6 Iron fortification... 6 Double emulsions for encapsulation... 8 Physical stability of W1/O/W2 double emulsions... 9 Double emulsion composition Vegetable oils Chemical stability of double emulsions: lipid oxidation Antioxidants Meat analogues Protein oxidation Material and Methods Materials Methods Preparation of W1/O/W2 double emulsions Physical characterization of double emulsions Chemical characterization of double emulsions: lipid oxidation Chemical stability of fortified meat analogues with iron-loaded double emulsions Statistical analysis Result and Discussion Physical stability of W1/O/W2 double emulsions Droplet size distribution Creaming

4 Morphology Encapsulation Efficiency (EE%) Chemical stability Primary lipid oxidation compounds Secondary lipid oxidation compounds Effect of antioxidant (α- tocopherol) Chemical stability of fortified meat analogues with iron-loaded double emulsions 35 6 Conclusion Recommendations References Appendix Appendix I Encapsulation efficiency I.I Calibration curves Appendix II Encapsulation efficiency of coconut oil and olive oil blends W1/O/W

5 1 Introduction Nowadays, as a consequence of the focus towards sustainable food production and animal welfare, there is a trend to produce meat substitutes based on plant protein (Ruby, 2012). However, meat has a very distinct and complex structure. To imitate both physical and chemical properties as well as the nutritional value of meat in the meat analogue has become a challenge for food researchers. Meat is rich in relatively high bioavailable iron form (heme iron), of which only 15-35% is absorbed compared to lower than 10% from plant origin (nonheme iron) (Ward et al. 2016). Iron is crucial to biologic functions in human being including respiration, energy production, DNA synthesis and cell proliferation (Camaschella, 2015). It is reported that iron deficiency has affected over two billions people in the world and it remains the most prevalent cause of anemia worldwide (Camaschella, 2015). Considering the low level of the bioavailable iron intake, for instance, vegetarians, children and pregnant women might be most vulnerable to iron deficiency which causes anemia and even cognitive dysfunction (Hunt, 2003) (Miller et al. 2013). Iron deficiency could be effectively alleviated by food fortification with iron (Ward et al. 2016). Therefore, iron fortification of a meat analogue is currently being investigated. In an order of preference of iron compounds for fortification, ferrous sulfate (FeSO4) is the first option due to high bioavailability and low cost (Zimmermann et al. 2007). However, FeSO4 accelerates lipid and protein oxidation which can cause undesirable defects in food products (Berton-Carabin et al. 2014). Because of its oxidative capacity and an unpleasant metallic taste, iron fortification is a challenge (Bittencourt et al. 2013). In order to conquer the downsides of the iron fortification, water-in-oil-in-water (W1/O/W2) double emulsions is a strategy to encapsulate hydrophilic bioactive iron like FeSO4 (Sapei et al. 2012). Entrapping iron in the inner water phase in (W1/O/W2) cannot only impede Fe 2+ to be released and oxidized to less bioavailable Fe 3+, but also mask the metallic taste. However, the drawback of the double emulsions for iron encapsulation are iron release due to the instability of double emulsions and lipid oxidation that can occur at W/O interface. The previous study of Eefje (2016) indicated that double emulsions prepared by sunflower oil showed a noticeable lipid oxidation after one day storage and poor physical stability which had 40 mmol/kg oil of conjugated dienes after 7 days. To avoid lipid oxidation and improve stability, Willem (2017) used coconut oil as the oil phase due to the with high saturated fat content. The formation of fat crystals increased encapsulation efficiency of 84.7% after 7 days. Nevertheless, when it comes to health aspect, the unsaturated oil is still preferable to design new food products. Hence, this study focused on designing double emulsions which provides better physiochemical stability and investigation of injection of double emulsions to meat analogue. 5

6 2 Objective The study aim was to firstly improve physiochemical stability of double emulsions for ironencapsulation in terms of oil types. Secondly, it was investigated the effect of α-tocopherol with different concentrations (100 ppm, 200 ppm and 300 ppm) to reduce lipid oxidation in olive oil double emulsions. Thirdly, it was investigated the chemical stability of meat analogues fortified with iron-loaded double emulsions by an injection method. 3 Theoretical background Iron Iron is an essential mineral that is part of the red blood cells constitution. Iron is crucial to biologic functions in human being including respiration, energy production, DNA synthesis and cell proliferation (Camaschella, 2015). Besides, iron is also involved in redox processes in the human body (Martınez-Navarrete et al. 2002). The biological importance of iron exhibited at its capacity of existing in several oxidation states. Ferrous (Fe 2+ ) and ferric (Fe 3+ ) are the predominant forms in the body, although higher-valent Fe 4+ or Fe 5+ is generated during the catalytic cycle of a number of enzymes such as catalases, peroxidases, and cytochromes P450. The total amount of iron in the human body is 4 grams (Martins, 2012). Around only 1-2 mg of iron is needed to replace the daily losses due to such as shedding of cells from intestine, skin and urinary tract or menstruation. In food, iron is present as heme iron (30-70%) and non-heme iron in meat and only as nonheme iron in plant. Non-heme iron has a low bioavailability which less than 10% is absorbed in the intestine (Ward et al. 2016). Since non-heme iron has a low bioavailability, people with low iron intake are more prone to iron deficiency. Iron deficiency may lead to unusual tiredness, shortness of breath, a decrease in physical performance, cognitive dysfunction and may influence the immune system in children and adults (Martınez-Navarrete et al. 2002). There are over two billions people in the world suffering iron deficiency and it is the most common cause of anaemia worldwide (Camaschella, 2015). However, vegetarians may be at particular risk from iron deficiency attributed by plant-based diets, containing low levels of poorly bioavailable iron (Ward et al. 2016). Therefore, the approach to increase iron intake among the population is to use iron supplements and increase the intake of iron-rich food. Iron fortification The fortification of iron in food is often regarded as the most cost-effective, long-term approach to ameliorate prevalence of iron deficiency. Iron fortified foods such as targeted wheat flour, cereals, pasta, milk, salt, beverage and condiments are widely consumed by at-risk iron deficient populations (Hurrell, 2007). To fortify the meat analogue such as vegetarian meats 6

7 will be a new step for food researchers. When developing an effective iron fortified food product, few prerequisites should be taken into account regarding choosing the iron compound (Martınez-Navarrete, 2002): (1) Iron compound should be water soluble and provides high bioavailability in human body. (2) Metallic taste of some iron compounds should be masked (3) Iron compound should be resistant to oxidation and discoloration. (4) Food vehicles should be well absorbed to the body and do not interfere in iron absorption. (5) A combination of an iron compound and food vehicle must be selected which is safe, acceptable to and consumed by the target population. Different forms of iron have been used to fortify different food products. Some characteristics of commonly used iron compounds are shown in Table 1. Table 1 Characteristics of iron compounds for food fortification water soluble Ferrous sulfate 7H2O Dried ferrous sulfate Ferrous gluconate Ferrous lactate Ferric ammonium citrate Approximate content (%) Fe Average bioavailability a relarelative cost (per mg/fe) b Poorly water soluble/soluble in diluted acid Ferrous fumarate Ferrous succinate Ferric saccharate Water-insoluble/poorly soluble in dilute acid Ferric orthophosphate Ferric ammonium orthophosphate (EKA Nobel, Sweden) Ferric pyrophosphate Elemental Fe powders: electrolytic carbonyl H-reduced Protected compounds NaFe EDTA Hemoglobin Adapted from (Hurrell, 1997)

8 a relative to hydrated ferrous sulfate (FeSO4.7H2O), in adult humans. b Relative to hydrated ferrous sulfate = 1.0, for the same level of total iron The water soluble iron compounds have the highest relative bioavailability among all the iron fortificants, therefore it is often the preferred choice. Ferrous sulphate is by far the most widely used and least expensive water soluble iron fortificant. The water soluble forms of iron are often used for cereal flours and dry foods like pasta, milk powder. However, these compounds can easily contact and react with the oxygen in the water or in the food matrix and lead to oxidations. Therefore, encapsulated iron which are physically separated from the food matrix can be used as an approach to avoid and slow down reactions that lead to unacceptable sensory changes in the food matrix (Allen et al. 2006). Double emulsions for encapsulation Encapsulation of foods compounds have been extensively applied in food industry for different purposes, for instance protecting materials from moisture, heat or other extreme conditions as well as mask odours or taste. There are various techniques being exploited to form the capsules, including spray drying, spray cooling, extrusion coating, fluidized bed coating, liposome entrapment (emulsions), coacervation and etc (F. Gibbs et. al. 2009). The controlled release, carrier material and preparation methods are the main factors concerned in food encapsulation. The double compartment structure of double emulsions gives an increasing interest, as they can be considered as an entrapping reservoirs for active substances to be released under a controlled transport mechanism (Benichou et al. 2004). W1/O/W2 double emulsions are used to entrap and protect the hydrophilic bioactive compounds which is encapsulated in W1 under a mild process. This may provide some advantages for iron fortification including masking undesirable flavours, preventing oxidation in the products, controlling the release of iron compounds and protecting labile ingredients from decomposing during eating and digestion (Jiménez, 2013). Thus, W1/O/W2 double emulsions is a potential strategy to compromise the drawbacks of iron fortification such as poor oxidative stability of iron form and metallic taste. Double emulsions are emulsions of emulsions which means the droplets of one dispersed liquid are further dispersed in another liquid. There are two types of double emulsions systems, water-in-oil-in water (W1/O/W2) and oil-in-water-in-oil (O1/W/O2) (Benichou et al. 2004). Here we mainly focus on W1/O/W2 double emulsions because water soluble iron compound (ferrous sulphate) which is encapsulated in internal aqueous phase (W1) is applied (Figure 1). However, double emulsions are an unstable system. 8

9 Figure 1 Water-in-oil-in-water double emulsions for iron encapsulation. Adapted from (Jiménez, 2013). Physical stability of W1/O/W2 double emulsions In emulsions, when two phases are in contact with each other, the molecules at the interface experience an imbalance of forces. Double emulsions are thermodynamically unstable systems because of the excess free energy at the interface between two phases of different polarity. The larger the interface, the higher the interfacial free energy and the more unstable system formed (Graber, 2010). The system tends to reduce the interface by various destabilizing mechanisms eventually leading to a complete phase separation (Figure 2) (Graber, 2010). Figure 2 Destabilization mechanisms in (W1/O/W2) emulsions. Source from (Graber, 2010). 9

10 The emulsifiers such as PGPR and Tween 20 are therefore added to W/O and O/W interface to reduce the instability of the emulsions. Destabilizing mechanisms such as coalescence and loss of inner water droplets (W1) due to diffusion to W2 are known as the main mechanisms of encapsulate compounds release in W1/O/W2 double emulsions (Graber, 2010). Coalescence may occur between two or more oil droplets, between two or more inner water droplets, or between inner droplets and external water phase. Another case is the rupture of the oil layer which consequentially causes the inner water droplets release to the external water phase. These mechanisms are shown in Figure 2 (Graber, 2010). Nevertheless, the stability of double emulsions also depends on various factors such as the particle sizes of W1, the viscosity of the emulsion, the density of the oil and the density of the aqueous phase (W1 and W2) according to Stokes Law. Hence the composition of the double emulsions plays an important role in the physical stability. Double emulsion composition Moreover, the composition of the double emulsions also affects their oxidative stability. And it mainly depends on the encapsulated compound for instance encapsulating iron which is a prooxidant and the oil type will be an important factor when the emulsion is designed. Vegetable oils Vegetable oils contain glycerol molecules esterified with different fatty acids, triacylglycerols (TAG), depending on plant patterns. Fatty acids (FA) can be classified in groups as saturated, monounsaturated (MUFA) and polyunsaturated FA (PUFA). Both physical and chemical features of oils are greatly influenced by the composition of the FA. In general, oils with high level of unsaturation are more susceptible to lipid oxidation. However, the different types of unsaturated FA also play an important role in oxidative stability of oils (Zambiazi et al. 2007). In vegetable oils, typically, olive oil with abundant MUFA and sunflower oil with high level of PUFA give an interest in the investigation of oxidative stability in double emulsions. The fatty acid composition of olive oil and sunflower oil is showed in Figure 3. The molecule formula of three most abundant fatty acids in both olive oil and sunflower oil are present in Figure 4. 10

11 linolenic linoleic oleic palmitic Content (%) Olive oil Sunflower oil Figure 3 Fatty acids composition of olive oil and sunflower oil (data from Grompone, 2011) (a) (b) (c) Figure 4 (a) molecule formula of palmitic acid (16:0), (b) molecule formula of oleic acid (18:1), (c) molecule formula of linoleic acid (18:2) Olive oil predominantly contains triacylglcerols (~99%) and a small proportion of secondarily free fatty acids, mono- and diacylglycerols, hydrocarbons, sterols, aliphatic alcohols, tocopherols and pigments. Besides, there are also excess of phenolic and volatile compounds naturally present in the oil. Moreover, olive oil has a high level of MUFA and antioxidants that contribute to a relatively oxidative stability. Four major FA presents in olive oil are oleic ( %), palmitic ( %), linoleic ( %), linolenic ( %) (Boskou et al. 2006). Furthermore, olive oil consists of substantial amount of phenolic compounds (>500 ppm), e.g., tyrosol, hydroxy tyrosol, and caffeic acid as well as total tocopherols especially a-tocopherol 11

12 (~ ppm) (Kamal-Eldin et al. 2007; Boskou et al. 2006). These compounds are powerful inhibitors in lipid peroxidation functioning as chain-breaking and metal ion-chelating agents (Kamal-Eldin et al. 2007). Hence, using olive oil as the oil phase in the double emulsions has a big advantage on preventing lipid oxidation. Similar to olive oil and most vegetable oils, sunflower oil primarily consists of triacylglycerols (98-99%) and a small amount of phospholipids as well as unsaponifiable matter including tocoperols, sterols and waxes (Chowdhury et al. 2007; Grompone, 2011). Different from olive oil, sunflower oil contains very high level of PUFA which is more prone to lipid oxidation (McClements et al. 2000). Four major fatty acids compositions of sunflower oil and average contents are linoleic (56.5%). oleic (33.7%), palmitic (5.2%), stearic (3.7%). Besides, sunflower oil contains a considerable amount of tocopherols, in which a-tocopherol has a level of 671 ppm (Grompone, 2011). Chemical stability of double emulsions: lipid oxidation Oxidation of oils during processing and storage is inevitable. Temperature, light, oxygen concentration, the presence of transition metals and fatty acid composition affect the oxidative stability of oils via autoxidation and photosensitized oxidation. Oxidative stability is an important indicator to determine oil quality and shelf life in terms of a sensorial alteration and formation of off-flavour compounds. Lipid oxidation also reduces essential fatty acids and produces toxic compounds and oxidized polymers which greatly influences the nutritional quality of the oil contained products (Choe et al. 2006). With regard to autoxidation of oils, it includes three steps, initiation, propagation and termination. See Figure 5. Figure 5 Interaction of antioxidants in lipid oxidation processes. Adapted from (Huss, H. H, 1995). 12

13 Free radical forms of fatty acids (L) are produced by removing the hydrogen atom from fatty acids (LH) in the initiation step which can be accelerated by heat, light and metal calalysts. The lipid radical (L) reacts quickly with atmospheric oxygen generating peroxy-radical (LOO) which may abstract a hydrogen from another LH resulting in a hydroperoxide (LOOH) and a new radical (Huss. H, 1995) ( Figure 5) The double bond near carbon radical in linoleic acid tends to shift to the more stable carbon position and cis form is rearranged to trans form. The primary oxidation products of linoleic and linolenic acids are mainly conjugated dienes hydroperoxides whereas peroxides are produced by oleic acids shown in Figure 5) (a) (b). Figure 6 (a) formation of conjugated dienes hydroperoxides from linoleic and linolenic acids (b) formation of hydroperoxides from oleic acids (Shanta et al. 1994). Some hydroperoxide positional isomers are also formed from oleic, linoleic and linoleic acids. The primary oxidation products are relative stable at room temperature when the transition metals are absent. However, high temperature or transition metals present accelerates the decomposition of hydroperoxides to alkoxy which can be further transformed to aldehydes, ketones, acids, esters, alcohols, and short-chain hydrocarbons via homolytic cleavage between oxygen-oxygen bonds (Choe et al. 2006). The time for the formation of secondary products from primary products differs from different oils due to their fatty acid compositions. In olive oil and rapeseed oils, secondary oxidation compounds are formed immediately after hydroperoxide formation as they are rich in monounsaturated fatty acids. However, secondary oxidation products are formed when the concentration of hydroperoxides reaches a considerable level in sunflower and safflower oils which contain plentiful polyunsaturated fatty acids (Choe et al. 2006). In W1/O/W2 double emulsions with unsaturated vegetable oil as the oil phase, lipid oxidation can occur promptly as the interfaces of oil and water provide large surface area that facilitates interactions between the lipids and water-soluble prooxidant compounds. Moreover, other factors can also potentially affect the oxidative stability of double emulsions such as fatty acid composition, concentration of antioxidants and prooxidants, lipid droplet characteristics (thickness, charge, rheology and permeability) (Waraho et al. 2011). 13

14 Futhermore, metals also accelerate lipid oxidation by decomposing lipid hydroperoxides. Fe 2+ reacts much faster than Fe 3+, however, Fe 2+ also causes the decomposition of phenolic compounds such as caffeic acid in olive oil and decreases the oil oxidative stability (Choe et al. 2006). The mechanism of decomposition of hydroperoxides by metal catalysts is shown in Figure 7. Figure 7 Decomposition of hydroperoxides by metal catalysts Antioxidants Vegetable oils originally contain antioxidants such as tocopherols, tocotrienols, carotenoids, phenolic compounds and sterols (Choe et al. 2006). Some of these antioxidants can have a synergistic effect. The types and content of the antioxidants vary from different oils. Antioxidants are compounds which retard lipid oxidation by a variety of different mechanisms including control of oxidation substrates (oxygen and lipids), control of pro-oxidants (reactive oxygen species and pro-oxidant metals), and inactivation of free radicals (McClements et al. 2000). A combination of metal chelator and free radical-scavenging antioxidants gives oil a better oxidative stability since they react in initial step and propagation step of lipid oxidation respectively (Choe et al. 2006). The metal chelator can bind the transition metals by chemical bonds and reduce the mobile transition metals which promote oxidative damage. And the interaction of antioxidant by scavenging radicals during lipid oxidation is demonstrated in Figure 5. Antioxidants like ascorbic acid, caffeic acid, EDTA and α-tocopherol are commonly used in food products. Ascorbic acid can act as a prooxidant oxidized by Fe 2+ and facilitate the lipid oxidation. EDTA can bind the iron significantly and inhibit the oxidation but it gives unpleasant flavour to food products. As for caffeic acid and α-tocopherol, both of them can inhibit the initial reaction between lipid hydroperoxides and Fe 2+ to a similar degree relating to their free radical scavenging and metal chelating activities (Aaneby. J, 2012). Nevertheless, caffeic acid can only dissolve in ethanol which is difficult to be added in the emulsions. Hence, α-tocopherol seemed a best choice among them. The antioxidant α- tocopherol belongs to tocopherol group which is the most important natural antioxidants presenting in the edible oils in forms of α-, β-, γ-, δ- tocopherol. And α- tocopherol is regarded as the most abundant and biologically active tocopherol in food ( Hraš, A. R, et al. 2000). As an antioxidant, α- tocopherol retards the lipid oxidation by scavenging the free radicals generated during the initial and propagation steps (McClements et al. 2000). 14

15 Meat analogues There are many studies about how to mimic meat to give a desirable structure to meat analogue in the Food Process and Engineering department in Wageningen University. However, there is a lack of research about the nutritional profile in meat analogue. Meat analogues for instance soy meat lack some important compounds for human body compared to meats such as iron, vitamin B12, Zinc and etc. Concerning the prevalence of iron deficiency and the trend of reducing meat consumption regarding the environmental aspect, the iron fortification is important to be investigated in the meat analogue (Camaschella, 2015). The applications of the fortification of iron in meat analogue are rather rare due to the technical barriers such as the bioavailability, enhancers/ inhibitors, product stability, sensory acceptability and safety (Uauy, R., et al. 2001) Thus, the double emulsions are designed to encapsulate iron compounds during the processing and applied to the meat analogue. Protein oxidation The meat analogue used in this study mainly consists of soy protein and the addition of double emulsions with iron compounds might lead to protein oxidation. Besides lipid oxidation, protein oxidation is also one of the substantial factors which influences the shelf life of food products. The most common damage for oxidized protein is the formation of carbonyls. There are four major pathways to generate carbonyls which include (Zhang, W, et al. 2013): (a) Fragmentation of backbones through α- amidation pathway and β-scission. (b) Binding of non-protein carbonyl compounds from lipid peroxidation by Michael addition (4-hydroxy-2-nonenal (HNE) and malondialdehyde) to protein amino acid side chains including histidine imidazole (P-His), cysteine sulfhydryl (P-SH), and lysine amino groups (P-NH2) as shown in Figure 8. Figure 8.Carbonyl derivatives produced by reaction with lipid peroxidation product (HNE) (Stadtman, E. R. et al. 2000). 15

16 (c) Direct oxidation of amino acid side chains including arginine, lysine, proline, and threonine (d) Addition of reactive carbonyl derivatives (ketoamines, ketoaldehydes, and deoxyosones) generated by reducing sugars and their oxidation products after reacting with lysine. 4 Material and Methods Materials Extra virgin olive oil (Jumbo, L-56 B, Lot), sunflower oil (Jumbo, , Lot) were used as the oil phase. Polyoxyethylenesorbitan monolaurate (Tween 20) (Sigma-Aldrich, USA) and polyglycerol polyricinoleate (PGPR) (Quest International, The Netherlands) were used as hydrophilic and lipophilic emulsifier respectively. D (+)-Glucose monohydrate (BDH Prolabo, VWR International) was added in W2 phase to counterbalance the osmotic pressure exerted by iron in the W1 phase. Deionized water (MilliQ, Merck Millipore) and ferrous sulphate heptahydrate (EMPROVE, Merck Millipore) were used for the preparation of the inner aqueous phase. BaCl2.2H2O (Sigma, B-0750), NH4SCN (Sigma, A-0302), HCL (4N)( AVS TITRINORM, Lot 17C134044), n-hexane (PEG GRADE. Actu-ALL Chemicals UN 12080), isopropanol ( Lichrosolv, Lot K ), methanol (Sigma aldrich Lot#STBG9740), 1-butanol (Sigma aldrich, Lot #BCBP6378V), para-anisidine (Aldrich, Lot #BCBN9155V), acetic acid (aldrich, J9270V. Lot; BCB) were used for the lipid oxidation measurements. Diaminoethane tetraacetic acid (EDTA) (Sigma Lot BCBQ 0232V), Sodium chloride (Sigma S9888/ Lot SZBC 0050V), 2,4 dinitrophenylhydrazine 97% (D199303/Lot MKBM2243V Aldrich), Tris (Tris (hydroxymethyl) aminomethane) (Sigma Lot BCBQ0063V), Trichloroacetic acid (Sigma 27242/ Lot STBG 4120V), Sodium Dodecil Sulfate (Sigma SLBQ0605U), Guanidine Hydrochloride (Sigma 50940/ Lot BCBN5507V), Ethanol (Emsure, ACS 99% ) and Ethyl Acetate (Emsure, ACS, ) were used for the protein extraction and protein oxidation measurements. Soy protein concentrate (SPC) - Alpha 6 ZP (Solae, St Louis, MO, USA) was used to make the meat analogues. Reagent A consists of bicinchoninic acid, sodium carbonate, sodium tartrate, and sodium bicarbonate in 0.1 N NaOH (final ph 11.25) and reagent B consists of 4% (w/v) copper(ii) sulfate pentahydrate were used to do BCA test. 16

17 Methods Preparation of W1/O/W2 double emulsions The primary emulsion (W1/O) was prepared with 25% (v/v) aqueous phase and 75% (v/v) oil phase using a rotor-stator homogeniser (IKA T18 basic Ultra Turrax, Staufen, Germany) for 4 min at a speed of rpm. The aqueous phase was added into the oil phase gradually with pipette during the emulsification process. The aqueous phase contained 0.5M FeSO4.7H2O and the oil phase was prepared by homogenizing 5 wt% PGPR with sunflower oil or olive oil using a magnetic stirrer (IKA RCT basic) at room temperature for 1 hour at 200 rpm speed. The double emulsions were prepared with gradually dispersing 20% (v/v) primary emulsion (W1/O) to 80% external water phase (W2) using a rotor-stator homogeniser (IKA T18 basic Ultra Turrax, Staufen, Germany) for 2 min at speed of rpm. The external water phase (W2) consisted of 2 M glucose monohydrate and 0.5 wt. % Tween 20. The glucose monohydrate was firstly dissolved in the water using a magnetic stirrer at 35 C for one hour at the speed 500 rpm. After cooling down the glucose solution, the water soluble emulsifier 0.5 wt. % Tween 20 was added to the glucose solution and mixed by a magnet stirrer at room temperature for 1 hour. The double emulsions (W1/O/W2) were stored at room temperature in the dark for the experiments of different time slots. Physical characterization of double emulsions Droplets size distribution The droplets size distribution of the double emulsions were analyzed with a static light scattering analyzer (Mastersizer 3000, Malvern Instruments Ltd, Worestershire, UK). The samples were stirred by a plastic pipette before being dispersed into MilliQ water at speed of 1200 rpm until an obscuration rate from 10% to 15 % was obtained. The refractive index used for sunflower oil was 1.465, for olive oil was and for the dispersant water was The measurements were carried out in fresh samples and after 1 day, 7 days and 14 days of storage. All the samples were measured for 2 times and each measurement was repeated for 3 times. Creaming index Creaming index was calculated to estimate the double emulsions stability during storage. Each sample of 10ml was poured into a 15 ml tube and stored at room temperature (± 20 C) in the dark to measure the creaming index (CI) in fresh (0h) and after 1 day, 7 days and 14 days of storage. The total emulsion volume (Vtotal) and the serum layer volume (Vserum) was measured over time and the CI was calculated as shown in Eq 1. 17

18 CI (%) = V serum V total 100 [Eq. 1] Morphology Light microscopy (Axio Scope.A1 Carl Zeiss Microscopy GmbH, Germany) was used to investigate the W1/O/W2 double emulsion morphology. The samples were gently stirred by a plastic pipette before analysis to ensure emulsion homogeneity. A drop of the emulsion sample was placed on a microscope slide and covered with a glass coverslip. The microstructure of double emulsions was characterized at room temperature (± 20 C). The images were captured (AxioCam MRc 5 camera) at a 40x magnification. Encapsulation efficiency The encapsulation efficiency (EE) was determined by iron concentration in the external aqueous phase (W2) through conductivity measurement (Hach, HQ14d, USA). The conductivity values were converted into iron concentrations using a calibration curve (Appendix I). The calibration curve was determined by the concentrations of iron dissolved in the W2 phase to exclude the conductivity values of the W2. EE % was defined as the percentage of iron still entrapped within the inner aqueous phase (W1) and calculated with the ratio of iron released into W2 at a specific time (Mt) relative to the total amount present in the external aqueous phase if all iron have been released from the W1 (M ). EE% = (1 M t M ) 100 [Eq. 2] M was set at M for double emulsions containing 20% (v/v) primary emulsion. Chemical characterization of double emulsions: lipid oxidation The primary lipid oxidation products of olive oil and sunflower oil were measured by peroxide value and conjugated dienes, respectively. The secondary lipid oxidation products were determined by para-anisidine value (p-av). 18

19 Peroxide value (PV) Doubled emulsion of 0.3 g was weighed in a 2 ml Eppendorf tube. The lipid fraction was extracted with 1.5 ml hexane: isopropanol (3:1 v/v). The solution was vortexed 3 times for 10 seconds with 20 seconds intervals and then centrifuged for 2 min at max speed. The supernatant was obtained for the assay. The calibration curve was prepared using cumene peroxide with a range of concentration of µm. The reagent was prepared by firstly mixing 0.9 ml of FeSO4 (0.04g/ 1 ml miliq water) with 0.9 ml BaCl2 (0.8g/ 25 ml 0.4M HCl) following a centrifugation and then mixing the supernatant with 0.9 ml of NH4SCN (7.5 g/25 ml miliq water). Then, 0.10 ml of the upper organic layer of extractions, calibration solutions and miliq water were mixed with 1.4ml of methanol: butanol (2:1 v/v) and 15 µl SCN / Fe 2+ (assay reagent). The tubes were covered and vortexed. The absorbance was read at 510nm after 20 min. If an absorbance was out of , it needed to be diluted and the result was adjusted to the dilute factor. The peroxide value was calculated by Eq.3 and the result is expressed in mmol/ kg oil. PV(mm kg oil) = C in oil+solvent Oil fraction (%) [ Eq.3] C in oil + solvent is the final peroxide concentration in oil and hexane (µm). Oil fraction (%) is the oil fraction of oil content of emulsion in hexane phase. Conjugated dienes hydroperoxide Firstly, 50 µl of double emulsions were added to 950 µl isopropanol. The blanks were prepared in the same way with 50 µl miliq water. The solutions were vortexed for 5 seconds at 2500 rpm 2 times. Then, 100 µl of the samples and blanks were added to 900 µl of isopropanol. The solutions were vortexed as above described. The supernatant was collected after centrifuging for 4 min at 1200 xg. The absorbance of the last supernatant was read at 233 nm and the absorbance spectra between nm was recorded. The concentration of conjugated dines the hydroperoxides (CD) was calculated using following formula: Concentration CD ( mol ) = A 233 [Eq.4] l Ɛ Ɩ A233 is the absorbance at 233 nm, Ɛ is the molar extinction coefficient at 233 nm ( = M - 1 cm -1 ) and Ɩ is the optical length in cm -1 ( =1). 19

20 Para- anisidine value Firstly, 2.5 ml of hexane/ isopropanol (3:1 v/v) was added to 1 g of double emulsions. Then, the solutions were vortexed 3 times for 10 seconds at 2500 rpm. The hexane phase was separated and then centrifuged at 1200 x g for 2 min. The hexane phase absorbance (Ab) was read at 350 nm and pure hexane was used as a blank. Then, 0.1 ml of acetic solution of para- Anisidine (2.5 g/l) was added to 0.5 ml of samples and pure hexane (blank). The solutions were vortexed for 10 seconds and then the absorbance of the hexane phase with para-anisidine (As) was measured at 350 nm. The results of the pav value were analysed in One-way analysis of variance(anova) from different groups at 0.05 significant level. The pav (arbitrary units) was determined as followed: m = mass (g) of oil per ml hexane. pav = (1.2As Ab) m [Eq.5] Chemical stability of fortified meat analogues with iron-loaded double emulsions The chemical stability of meat analogues fortified with iron-loaded double emulsions was investigated with measuring the carbonyl content in samples during storage time. Here it will be described how the meat analogues were prepared, the injection of the iron-loaded double emulsions and later the carbonyl content measurement. Meat analogues preparation The meat analogues were prepared in a High Temperature Shear Cell (HTSC) (Wageningen University) using soy protein concentration (SPC) as the protein matrix. Firstly, 54wt% demi water and 1% of NaCl were mixed. Then, 45wt% SPC was added to the solution and mixed before the matrix was left for hydration for 30 min at room temperature. After 30 min, the matrix was put in HTSC for 15 min at 30 rpm at 140 C. After 15 min, there was a cooling down for 5 min at 25 C. Afterwards, the samples were stored in plastic bags and frozen for further analysis. Injection of iron-loaded double emulsions The injection of iron-loaded double emulsions was done with a syringe (0.5 mm*16 mm). According to the common level of iron fortification in different food products (10 ~ 80 mg Fe per 100 g product), it was decided to inject ml double emulsions per 1 g of meat analogue 20

21 (40 mg Fe per 100 g product) (Martınez et al. 2002). The meat analogues samples were approximately divided in 16 equal areas and the same amount of double emulsion was injected to each area (0.014 ml/g). After the injection, each sample was cut into 3 portions (~23g per portion) for the measurements at 0 h, after 1 day and 7 days of storage. The samples were stored in 2 plastic bag with vacuum at 35 C in the oven for the measurements at day 1 and day 7. Separation of protein fractions for DNPH method Meat analogues fortified with iron-loaded double emulsions and control samples were cut into small pieces using a punch with 8 mm size and placed in a box and mixed. Then, 9 g of samples were weighed in a 250 ml centrifuge bottle for protein fraction separation. Buffer 1 (1:3 w/v) (5 mm and 100 mm Tris ph 7.5) was added to the samples. The samples were then homogenized for 60s at rpm using an IKA Ultra-Turrax in an ice bath, followed by a centrifugation at x g at 2 C for 20 min. The supernatant was collected as the protein fraction 1. Then, buffer 2 (1:3 w/v) (5 mm EDTA, 100 mm Tris and 500 mm NaCl ph 7.5) was added to the pellet, which was disconnected with a spoon from the bottom to help homogenization process. Further, the sample was homogenized for 30s at rpm and centrifugation was done as previous described. The supernatant was collected as the protein fraction 2. To the resultant, pellet was added by 0.15 M KCL (1:3 w/v), following a homogenization for 30s at rpm in an ice bath using an IKA Ultra-Turrax. The protein separations were done for the measuring carbonyl content with DNPH method. And the protein content of DNPH samples were measured together in the end of the week by BCA assay. DNPH method Firstly, 1 ml of 20% (w/v) TCA was added to 1 ml of protein fraction 1 and 2. The solutions were centrifuged at x g for 5 min. After removing the supernatant, 400 µl of 5% (w/v) SDS was added to the pellet. The samples were heated at 99 C for 10 min in a thermomixer and then ultrasonicated at 40 C for 30 min at 80 KHZ and 90% sweep in a ultrasonicate bath After ultrasonication, 0.8 ml of 0.3% (w/v) DNPH in 3M HCl was added to samples and 0.8 ml of 3M HCl was added to the blanks. All the samples were incubated for 60 min at room temperature in the dark and vortexed for 5s at 2500 rpm every 10 min. After incubation time, 400 µl of 40% (w/v) TCA was added to precipitate the proteins, followed by a centrifugation at x g for 5 min. After, the supernatant was removed and the pellet was washed 3 times with 1ml of ethanol ethyl acetate (1:1, v: v) solution by centrifugation at 15,000 x g for 5 min each time. The final pellets were dissolved in 1.5 ml of 6 M guanidine hydrochloride (ph 6.5). After vortexing for 5 seconds, the samples were placed in a thermomixer overnight at 37 C for around 13h. The next morning, the samples absorbance were read at 370 nm. If the absorbance was out of the range 0.12 and 0.55, samples and blanks were diluted. 21

22 Carbonyl content was calculated by the following equation: using Abs sample Absblank = C ε mmol carbonyls ( kg ) (mmol) [Eq. 6] soluble protein concentration (kg) Where, Abs sample is the absorbance of the sample and Abs blank is the absorbance of the blank. The molar absorption coefficient of carbonyls at 370 nm is ε = M -1 cm -1. Soluble protein concentration: BCA assay BCA assay was used to measure the soluble protein content in protein fractions after the DNPH in the samples. The BCA working reagent was prepared by mixing 50 parts of reagent A with 1 part of B. The calibration curve was prepared in the range of 0 g/l to 1 g/l in guanidine or buffers for protein fractions Then,1 ml of working agent was added to 50 µl of samples and calibration curve and placed in a thermomixer for 30 min at 37 C and 300 rpm. After, the samples were cooled at 4 C for 5 min and after that at room temperature for 10 min. After this 10 min, the absorbance was read at 562 nm within 10 min. Statistical analysis Except for protein oxidation experiments, all the experiments were carried out in triplicates. Microsoft Excel software was used for data processing and statistical analysis. Mean values were compared with One-way analysis of variance(anova) from different groups at 0.05 significant level. 22

23 Volume Density (%) 5 Result and Discussion This section is divided into three parts. The first part covers the different influence of olive oil and sunflower oil on the physical stability of W1/O/W2. The second part discusses the chemical stability of olive oil W1/O/W2 and sunflower oil W1/O/W2 and the effect of different concentrations of α- tocopherol on lipid oxidation in olive oil W1/O/W2. The third part covers protein oxidation of injected meat analogue samples with W1/O/W2 double emulsions. Physical stability of W 1 /O/W 2 double emulsions Droplet size distribution The droplet size distribution of W1/O/W2 with sunflower oil and olive oil were measured in fresh samples and after 1 day, 7 days and 14 days of storage. Both W1/O/W2 prepared with sunflower oil and olive oil had considerably stable droplet size distribution over 14 days of storage shown in Figure 9 (A and B) (A) 0h 24h 7d 14d Droplet size (µm) 23

24 Volume Density (%) (B) Droplet size (µm) 0h 24h 7d 14d Figure 9 Droplet size distribution of W1/O/W2 emulsions prepared with sunflower oil (A) and olive oil (B) measured at fresh (0h) and after 1 day, 7 days and 14 days of storage. Results from one representative measurement. Thus, the double emulsions might be stable to coalescence and diffusion between W1 and W2, within two or more inner water droplets and the oil droplets containing W1. Different mechanisms of droplets coalescence and the diffusion between W1 and W2 phases were explained in We can conclude that this W1/O/W2 composition was well designed which provides the physical stability of W1/O/W2 prepared with sunflower oil and olive oil. Moreover, the overall droplet size distribution of W1/O/W2 with sunflower oil and that of W1/O/W2 with olive oil also demonstrated a similar trend and there was no significant difference in terms of average droplet sizes and span (p> 0.05) throughout 14 days of storage (See Table 2). Table 2 mean droplet diameter and span values of W1/O/W2 emulsions prepared with olive oil and sunflower oil (p>0.05). Results are expressed as mean values and standard deviation (n = 18). Olive oil W1/O/W2 Sunflower oil W1/O/W2 Time (days) D 4,3 (µm) Span D 4,3 (µm) Span Fresh 27.4 ± ± ± ± ± ± ± ±

25 Creaming index (%) Creaming The creaming index was measured in fresh W1/O/W2 emulsions with sunflower oil and olive oil and after 1 day, 7 days and 14 days. It can be seen that both double emulsions performed relative stable in creaming and showed similar results (Figure 10) Time (days) Sunflower oil W.O.W Olive oil W.O.W Figure 10. Creaming index of sunflower oil and olive oil double emulsions in fresh samples and after 1, 7 and 14 days of storage. Results are expressed as mean values and error bars are standard deviation (n = 3). The macroscopic appearances of W1/O/W2 with olive oil and sunflower oil during storage are displayed in Figure

26 a b c d e f g h Figure 11. Macroscopic appearance of W1/O/W2 emulsions prepared with olive oil and sunflower oil stored at room temperature. a, b, c, d are olive oil at time 0, after 1 day, 7days and 14 days of storage. e, f, g, h are sunflower oil at time 0, after 1 day, 7days and 14 days of storage 26

27 The creaming formed gradually but not evidently from 0 hour to 14 days of storage in both W1/O/W2 with olive oil and sunflower oil due to Ostwald ripening. A yellow layer on the top of the tube an undesired odour was noticed after 1day in sunflower oil W1/O/W2 and multiple yellow layers were formed after 7 days. Undesirable odour was noticed after 1 day. However, two yellow layers were only formed in olive oil W1/O/W2 after 14 days of storage and the undesired odour was generated at around 7 days. These findings suggested the presence of secondary lipid oxidation products. Lipid oxidation results will be discussed further. Morphology Microscopic images were taken over storage time (fresh, 1, 7 and 14 days) for both double emulsions (Figure 12). 27

28 Figure 12. Microscopic images of olive oil ( left column) and sunflower oil (right column) W1/O/W2 double emulsions from fresh samples (a,b) and after 1 day (c, d), 7 days (e, f) and 14 days (g, h) of storage. Scale bar represents 20 um (40x magnifiication) It was found that the droplet sizes from W1/O/W2 with olive oil (left) and sunflower oil (right) seemed similar and no obvious change in both during 14 days of storage which is in line with the result of droplet size distribution. However, we can see few empty oil droplets after 7 days of storage in both double emulsions, which means that the inner water phase was released to the external water phase (Figure 12). After 14 days, it seemed more noticeably but there was still marginal amount of empty droplets which could be also the droplets that were not focused under the microscopy. Hence, the release of W1 needs to be proven with encapsulation efficiency measurements. Encapsulation Efficiency (EE%) The EE% of sunflower oil W1/O/W2 and olive oil W1/O/W2 did not show a significant difference (p>0.05). The EE% had a slow decrease from 97% to 85% and 82% within 14 days in W1/O/W2 with sunflower oil and W1/O/W2 with olive oil respectively (Figure 13). 28

29 Encapsulation Efficiency (%) Sunflower Oil W.O.W Olive Oil W.O.W Time (day) Figure 13. Encapsulation efficiency (%) of sunflower oil W1/O/W2 and olive oil W1/O/W2. Results are expressed as mean and error bars are standard deviation (n = 3) The similar droplets size distribution between both sunflower oil W1/O/W2 and olive oil W1/O/W2 might result in a close encapsulation efficiency of both types of emulsions. This result was higher than it was found in previous study where only 57% of iron remained encapsulated after 7 days in double emulsions prepared with 15% sunflower oil phase (School, 2016). The lower oil phase content had lower EE% stability over time, which shows that higher oil phase content in double emulsion can considerably enhance the encapsulation efficiency. This could be explained that the higher oil content gives higher viscosity to the double emulsions which creates better stability according to Stokes law and hence improved the encapsulation efficiency of iron. Nevertheless, there is no study found about the influence of oil content in W1/O/W2. In another study, using coconut oil as oil phase obtained an EE% of 84.7% after 7 days of storage (Bax, 2017), which was commensurate to the result in the present study. Based on the results from the microscopy and EE%, we can conclude that the iron was encapsulated in the inner water phase efficiently in W1/O/W2 with both sunflower oil and olive oil during 14 days of storage. Hence, we can consider that the iron chelate might not be needed to control the iron release. Chemical stability The primary lipid oxidation compounds were measured by peroxide value for olive oil W1/O/W2 and conjugated dienes hydroperoxides for sunflower oil W1/O/W2. The secondary lipid oxidation compounds were measured by para-anisidine value for both W1/O/W2. Due to the inhomogeneity of the emulsion samples, the results were carried out with relative high standard deviation. 29

30 Peroxide Value (mmol/kg oil) Primary lipid oxidation compounds The peroxide value of olive oil W1/O/W2 was relatively low during 14 days of storage ( Figure 14), being stable in the first 24 hours and slightly increased after 7 days of storage. However, the peroxide value of olive oil W1/O/W2 dropped marginally after 14 days but it is still higher than the initial value in fresh samples Time (day) Figure 14. Peroxide value (mmol/kg oil) in fresh olive oil W1/O/W2 emulsions and after 1 day, 7 days and 14 days of storage. Results are expressed as mean and error bars as standard deviation (n =3). Overall, the peroxide value of olive oil W1/O/W2 did not show a significant change which could be attributed by the natural abundant of antioxidants present in the olive oil (Visioli et al. 1998). The propagation step of lipid oxidation continues until one of the radicals generated from initiation step is removed by the reaction with an antioxidant (Huss. H, 1995). Hence, because of the interaction of antioxidants on propagation step, hydroperoxides value was controlled which provided the relatively stable peroxide value (Choe et al. 2006). However, this result is hard to compare with the other researches because of the unique method that was used in this study. Peroxide value was usually determined by titration method and calculated in peroxide milliequivalent per kg oil (meq/kg oil) and the mmol peroxide per kg oil was expressed in this study with the method of calibration curve. The conversion between two values were not found and though we can see the values seemed relatively low and stable, we could not define an acceptable value in this case. Therefore, the results of peroxide value should be considered together with the result of secondary oxidation. The conjugated dienes hydroperoxides in sunflower oil W1/O/W2 emulsions increased rapidly (p<0.05) in 1 day and then demonstrated a slow increase after 7 days of storage followed by a fast decrease (p<0.05) after 14 days (Figure 17). 30