Comparison of Chemical and Enzymatic Interesteri fi cation of Fully Hydrogenated Soybean Oil and Walnut Oil to Produce a Fat Base with Adequate Nutritional and Physical Characteristics

Lipids are the highest energy source of the three macronutrients (carbohydrates, proteins and lipids). They also add fl avour, texture and satiety to foods. Slight diff erences in the fat produce signifi cant changes in the food. For instance, a fat that performs well in baked products will not work well enough in ice cream, producing a pasty and waxy feeling instead of a pleasant cooling eff ect (1). On the other hand, some fats may have excellent physical properties for specifi c applications but are not recommended due to nutritional concerns. These lipids may not be absorbed as expected or may even have some deleterious eff ects on health due to the presence of certain fatt y acids. Each of the three fatt y acids bonded to the sn1, sn-2 or sn-3 position of the glycerol backbone can vary in regard to chain length, number and position of double bonds, and geometrical confi guration. These characteristics give lipids their physical, chemical and nutritional properties (2,3). ISSN 1330-9862 scientifi c note


Introduction
Lipids are the highest energy source of the three macronutrients (carbohydrates, proteins and lipids).They also add fl avour, texture and satiety to foods.Slight diff erences in the fat produce signifi cant changes in the food.For instance, a fat that performs well in baked products will not work well enough in ice cream, producing a pasty and waxy feeling instead of a pleasant cooling eff ect (1).On the other hand, some fats may have excellent physical properties for specifi c applications but are not recommended due to nutritional concerns.These lipids may not be absorbed as expected or may even have some deleterious eff ects on health due to the presence of certain fatt y acids.Each of the three fatt y acids bonded to the sn-1, sn-2 or sn-3 position of the glycerol backbone can vary in regard to chain length, number and position of double bonds, and geometrical confi guration.These characteristics give lipids their physical, chemical and nutritional properties (2,3).The optimal physical, chemical and nutritional properties of an oil or fat are not always mutually compatible (4).As such, lipid modifi cation is a good way to give them specifi c functionalities, increase their oxidative stability, or improve their nutritional value.The temperature at which fats crystallize and melt or the ratio and kind of solid and liquid fats that produce a specifi c plastic behaviour are physical properties that are specifi c for each application (4).Concerns about trans confi guration or actions to reverse the obesity epidemic in which the aim is to reduce oil absorption stand in contrast to the need to increase the poor lipid absorption in an immature digestive system or in patients with cystic fi brosis (5,6).Our knowledge of lipid metabolism has played a part in the development of structured lipids.These lipids are obtained synthetically by changing the fatt y acid composition and/or their distribution in the glycerol backbone in order to meet a specifi c need and improve their nutritional or functional properties (7).Betapol TM and Salatrim are structured lipids produced with nutritional purposes.
One is a substitute for human milk fat and the other one is a reduced-calorie fat.Structured lipids may also mimic desirable physicochemical properties.Plastic fats and cocoa butt er equivalents prepared from lower value fats and oils can also be synthesized (8,9).
There are technological and biological methods of modifying oils and fats that may expand their uses.The biological methods include genetic engineering and crop or animal control (4).The technological methods include blending and fractionation, which are physical processes that produce value-added fats and oils (1).However, it is not always easy to predict what will happen to minor components that could aff ect the oxidative stability of products (4).Hydrogenation solved this problem, allowing for the production of fats with creaming properties, frying stability, sharp melting properties, or other functional characteristics for specifi c applications while enhancing oxidative stability (1).However, hydrogenated fats contain the two least desirable fatt y acids: saturated and trans (10,11).Full hydrogenation is an alternative that produces hard fats, which may be used to prepare low to zero-trans commercial fats through interesterifi cation (12,13).An interesting raw material that can be used in this process is fully hydrogenated soybean oil.This is a relatively low-cost product with a high C18:0 content (around 85 %).It is not atherogenic and has none of the adverse eff ects on cardiovascular diseases reported for shorter fatt y acids (C12:0, C14:0 and C16:0) (13).
Interesterifi cation is a reaction through which it is possible to rearrange the fatt y acids in the triacylglycerol molecule so that its composition changes but the fatt y acid profi le is preserved (14)(15)(16).There are two ways to carry out interesterifi cation: chemical and enzymatic.Chemical interesterifi cation requires an alkaline catalyst.It is relatively inexpensive, readily available, and easy to use and scale-up.However, it lacks specifi city, off ering litt le or no control over the position in which fatt y acids are distributed in the fi nal product (7,17).Enzymes can be used to increase the amount of control that one can have over the nature of the product as a consequence of the specifi city shown by many lipases.As such, enzymatical-ly interesterifi ed lipids have a more defi ned structure (11,18).The enzymatic process off ers milder reaction conditions and thus lowers degradation of long-chain polyunsaturated fatt y acids.Moreover, it also produces fewer by-products than the chemical process (19)(20)(21).In this respect, raw materials such as walnut oil are of great interest.Walnuts are unique within the nut family due to their high polyunsaturated fatt y acid content, specifi cally C18:3n3, and a ratio of C18:2n6/C18:3n3 of 4:1, which has shown to decrease the risk of heart disease (22).
In accordance, the aim of this study is to compare chemical and enzymatic interesterifi cation of binary blends of fully hydrogenated soybean oil and walnut oil and evaluate them in the synthesis of a zero trans-fat and high C18:3n3 fat bases.

Materials
Raw materials for chemical and enzymatic interesterifi cation were fully hydrogenated soybean oil supplied by Watt 's S.A. (Santiago, Chile) and walnuts donated by Valbifrut S.A. (Santiago, Chile).Walnut oil was obtained by cold pressing.The product was also neutralized and bleached (23).Fully hydrogenated soybean oil and walnut oil were stored at 4 °C in a nitrogen atmosphere until the experiments were conducted.Raw materials were chosen because of the saturated fatt y acid content of fully hydrogenated soybean oil and the associated high melting point (68-75 °C), and the high C18:3n3 content of walnut oil (14 %), as shown in Table 1.For chemical interesterifi cation, analytical grade sodium methoxide (95 %, Sigal Ltda., Santiago, Chile) and citric acid monohydrate

Chemical interesterifi cation
Blends of fully hydrogenated soybean oil and walnut oil made at three mass ratios (20:80, 40:60 and 60:40) were chemically interesterifi ed following the procedure described by Rodríguez et al. (24).Briefl y, each blend was dried under vacuum conditions (100 mm Hg) and heated with constant stirring (at 150 rpm) in a thermoregulated bath until it reached (90±2) °C.Next, 0.5 % (by mass) sodium methoxide was added and the reaction was carried out for 10, 15 or 60 min.Citric acid monohydrate was added to stop the reaction (1.78 g per g of sodium methoxide) and kept for 5 min.The interesterifi ed blend was then washed three times with distilled water to remove the produced soap and any residue of sodium methoxide or citric acid monohydrate.

Enzymatic interesterifi cation
Blends of fully hydrogeneated soybean oil and walnut oil made at three mass ratios (20:80, 40:60 and 60:40) were enzymatically interesterifi ed following the method reported by Abigor et al. (25).Each blend was dried under vacuum conditions (100 mmHg) and heated with constant stirring (at 150 rpm) in a thermoregulated bath until it reached (70±2) °C, a lower temperature than the one required for chemical interesterifi cation.Next, 5 % (by mass) Lipozyme TL IM was added.The reaction was carried out for 30, 120 or 240 min.In order to stop the reaction, the enzyme was removed by fi ltration.All of the chemically and enzymatically interesterifi ed blends were stored at 4 °C in a nitrogen atmosphere.

Solid fat content
Because interesterifi cation modifi es the melting profi le of lipids, which becomes constant when equilibrium is reached (10), the solid fat content was measured and used as an indicator of this.The solid fat content of interesterifi ed blends was measured using pulsed nuclear magnetic resonance (p-NMR) according to the AOCS Offi cial Method Cd 16-81 (26).Briefl y, dry and fi ltered samples were placed in glass tubes and completely melted (at 60 °C and 10 min) and then solidifi ed (at 0 °C and 30 min).Samples were then allowed to dissolve in water at 10.0, 21.1, 26.7, 33.3 and 40.0 °C for 15 min.Finally, the solid fat content was measured at each temperature in a Bruker Minispec PC120s p-NMR analyzer (Bruker Analytische Mestechnik, Rheinstett en, Germany).

Fatt y acid profi le
Methylated fatt y acids of the triacylglycerols were analyzed in an HP 5890 gas chromatograph (Hewlett --Packard, Palo Alto, CA, USA).A fused silica capillary column BPX70 (0.25 μm fi lm thickness, 50 m length, 0.33 mm i.d.; SGE Analytical Science, Austin, TX, USA) was used.Samples were run with hydrogen as the carrier gas between 160 and 230 °C at a rate of 2 °C/min.Standard fatt y acid methyl esters from Merck (Darmstadt, Germany) were used for identification purposes.

Statistical analysis
All of the analyses were carried out in triplicate and the results are expressed as mean values±standard error of the mean (SEM).Statistical diff erences between the times and mass ratios were determined using one-way analysis of variance and Fisher's test.Diff erences were considered signifi cant at p<0.05.Statistical analysis was performed using Statgraphics v. 4.0 (StatPoint, Inc., Warrenton, VA, USA).

Results and Discussion
Chemical and enzymatic interesterifi cation were carried out to obtain a fat base with a high content of C18:3n3 and zero trans-fats, as shown in Table 1.

Chemically interesterifi ed fat
Fig. 1 shows the solid fat content of blends of fully hydrogenated soybean oil and walnut oil with mass ratios of 20:80, 40:60 and 60:40 interesterifi ed for 0 (non-interesterifi ed), 10, 15 and 60 min.As we expected, the blends with higher fully hydrogenated soybean oil content had the highest solid fat content.Fully hydrogenated soybean oil is composed exclusively of saturated fatt y acids and may contain traces of unsaturated fatt y acids (Table 1).As the content of walnut oil in the blend increased, the solid fat content decreased due to the high content of unsaturated fatt y acids.The chemical interesterifi cation had the same eff ect on all blends: the reaction reduced the solid fat content at all temperatures.Our results echo those reported in other studies (15,24,27), of reduced melting point aft er interesterifi cation; however, the contrary eff ect -an increase of the solid fat content produced by interesterifi cation -may also happen (14,28).The chosen conditions, temperature and catalyst concentration, allowed thermodynamic equilibrium to be reached aft er a few minutes.This eff ect was refl ected in the stabilization of melting temperature.No signifi cant diff erences in solid fat content in the 20:80 blend were found with subsequent measurements aft er 10 min and in blends 40:60 and 60:40 aft er 15 min.Furthermore, the beginning of the reaction was retarded with increasing content of walnut oil.While the 20:80 blend had interesterifi ed in 10 min, the melting point of the 40:60 blend aft er 10 min of interesterifi cation was found at signifi cantly lower temperatures than at the beginning of the reaction, although it was still not stabilized.Interestingly, this particular blend (Fig. 1b) shows evident variability aft er 10 min (high standard error of the mean), probably denoting a transition state, which refl ects how unstable the solid fat content of a blend may be during the interesterifi cation.It can also be noticed that aft er 10 min of interesterifi cation, the melting point of the 60:40 blend did not change signifi cantly from that observed in the non-interesterifi ed blend.These results suggest that the beginning of the reaction is delayed with an increase in the content of fully hydrogenated soybean oil.Comparing our results to those reported, the reaction time may not be relevant because it might vary widely (5 min to 6 h, or even longer) depending on the conditions such as the catalyst concentration, temperature and solubility of catalysts in the reactants (15).In general, the reaction is accelerated with the increase in catalyst concentration and temperature, but the eff ect of solubility of the catalyst depends on the raw materials, mass ratios and the catalyst itself.

Enzymatically interesterifi ed fat
Fig. 2 shows the melting profi les of fully hydrogenated soybean oil and walnut oil blends at mass ratios of 20:80, 40:60 and 60:40 interesterifi ed for 0 (non-interesterifi ed), 30, 120 and 240 min.As we observed with chemical interesterifi cation, blends with higher content of fully hydrogenated soybean oil present higher melting tempera tures, which are reduced with the addition of walnut oil.Also, the solid fat content of all blends decreases as the reaction progresses.Díaz Gamboa and Gioelli (29) synthesized functional triacylglycerols using chemical and enzymatic interesterifi cation.They reported that the addition of a solid fat to liquid oil increases the solid fat content of t he blend, which decreases aft er interesterifi cation, as we found in our experiments.They also reported solid fat content at lower temperatures aft er interesterification of pure palm kernel fat.The thermodynamic equilibrium of all the fully hydrogenated soybean oil and walnut oil blends was achieved at 120 min when the solid fat content was stabilized.This coincides with the results of the study by Undurraga et al. (8), who reported total interesterifi cation aft er 80-120 min under similar conditions (65 °C, Lipozyme TL IM).However, as in chemical interesterifi cation, the temperature and the catalyst concentration aff ected the rate of the enzymatic reaction.Interesterifi cation speeds up when temperature increases until it reaches a maximum level.At that point, it starts to slow down as the temperature increases due to the loss of enzyme activity (30).The catalyst concentration generally has the same eff ect.As the enzyme load increases, the reaction rate is accelerated, but above a certain amount, there is no eff ect (31).

Interesterifi ed blends as a fat base for margarine
In order to obtain an interesterifi ed fat similar to a fat base for margarine, the melting profi les of a commercial fat base and interesterifi ed blends were compared.Fig. 3 shows the melting profi les of the interesterifi ed mixtures and the commercial base.Although diff erences between the curves of the commercial base and the studied mixtures are observed, the 40:60 blend shows physical behaviour similar to that of the commercial base.Chemically and enzymatically interesterifi ed blends exhibited a spreadability as good as a commercial fat with 27 % of solid fat at 10 °C; a solid fat content not greater than 32 % is essential for good spreadability at refrigeration temperature.Interesterifi ed blends also showed good stability and resistance to oil exudation at room temperature with a solid fat content over 10 % at 21.1 °C (13 and 16 % in chemically and enzymatically interesterifi ed blends, respectively).The poorest behaviour of interesterifi ed blends was at body temperature.Under these conditions, the contents of 9 and 11 % of solids at 33 °C for chemically and enzymatically obtained products produced a waxier sensation in the mouth.Ideally, the solid fat content should be less than 3.5 % at 33.3 °C (32-34).However, the solid fat content of both interesterifi ed blends had plasticity curves that fall within the range of all-purpose-type shortening fats.According to List et al. (35), the solid fat content required at 10, 21.1, 26.6, 33.3 and 40 °C is 18-23, 14-19, 13-14, 12-13 and 7-11 %, respectively.Overall, the interesterifi ed blends achieved these requirements.However, at 10 °C, the solid fat content was slightly higher than that of shortening fats.
There are no physical diff erences between the chemically and enzymatically interesterifi ed blends.Other differences between chemically and enzymatically interesterifi ed fats may be produced by the process itself.The main advantage of the enzymatic process over the chemical one is that the latt er produces complete randomization of fatt y acids, while enzymatic interesterifi cation may be either substrate specifi c, diff erentiated by chain length, or stereospecifi c.The stereospecifi city of commercial enzymes allows them to produce structured lipids with mainly nutritional advantages.Essential fatt y acids are usually bonded at the sn-2 position of triacylglycerol; if these oils are enzymatically interesterifi ed, the essential fatt y acids remain at the sn-2 position, where they are more easily absorbed as sn-2 monoacylglycerol than free fatt y acids (36).This improved absorption characteristic is not maximized in chemical interesterifi cation.Enzymatic interesterifi cation is mainly used when a very specifi c structure is required, as in Betapol ® , the natural standard for mimicking human milk.

Conclusions
This study evaluated the suitability of chemical and enzymatic interesterifi cation to produce a fat base rich in polyunsaturated fatt y acids.Specifi cally, fully hydrogenated soybean oil and walnut oil were shown to be adequate raw materials to produce a fat base with good physical and nutritional characteristics, with a high concentration of linolenic acid (C18:3n3), and with zero trans--fat.In addition, fully hydrogenated soybean oil may have additional advantages compared to other solid fats, due to the high concentration of palmitic acid (C18:0), which does not have as many detrimental eff ects as shorter saturated fatt y acids.Overall, both chemically and enzymatically interesterifi ed blends of fully hydrogenated soybean oil and walnut oil at 40:60 mass ratio resulted in the plasticity of the shortening fat and no signifi cant diff erences were found between both technologies.Additional studies related to the fatt y acid bioavailability of the diff erent mixes could be done in the future to determine if there could be any drawbacks in terms of fatt y acid availability when comparing a blend with its interesterifi ed mix.This analysis could also be reinforced through the study of the molecular distribution of the diff erent fatt y acids along the glycerol backbone aft er interesterifi cation.

Fig. 3 .
Fig. 3. Solid fat content (SFC) of chemically (--) and enzymatically interesterifi ed (--) blends of fully hydrogenated soybean oil and walnut oil at mass ratio of 40:60, and of a commercial fat base (-), measured at diff erent temperatures

Table 1 .
Fatt y acid composition of raw materials: fully hydrogenated soybean oil (FHSBO), walnut oil and FHSBO/walnut oil mixes