Investigation of the Effects of Inulin and β-Glucan on the Physical and Sensory Properties of Low-Fat Beef Burgers Containing Vegetable Oils : Optimisation of the Formulation Using D-Optimal Mixture Design

In recent years, excessive consumption of meat products has been associated with obesity, hypertension, ischaemic heart disease, even diff erent forms of cancer due to high amounts of saturated fatt y acids (SFAs), cholesterol and sodium (1,2). Since meat and meat products still play an important role in providing many essential nutrients for optimal growth, improving the quality of currently available products in order to develop new healthy meat products is essential.


Introduction
In recent years, excessive consumption of meat products has been associated with obesity, hypertension, ischaemic heart disease, even diff erent forms of cancer due to high amounts of saturated fatt y acids (SFAs), cholesterol and sodium (1,2).Since meat and meat products still play an important role in providing many essential nutrients for optimal growth, improving the quality of currently available products in order to develop new healthy meat products is essential.
In order to enhance the quality of meat products, a lot of research has been conducted recently to reduce the amount of animal fat and/or replace it with unsaturated fatt y acids in formulations for a variety of products (3)(4)(5)(6).Since vegetable oils contain higher levels of unsaturated fatt y acids and are free from cholesterol when compared to animal fats, their incorporation into meat formulations could improve nutritional benefi ts of the fi nal product (5).Reduction of animal fat is a big challenge for the meat industry because it causes negative alterations in the textural and sensory properties of the meat products (3,5,7).Numerous studies investigated possible addition of dietary fi bre, which could improve the textural and sensory properties but also enhance the nutritional properties.Therefore, inclusion of dietary fi bre is one of the main possible solutions currently investigated in food industry (3,8).
β-Glucan, which is known as a prebiotic, is a soluble dietary fi bre that originates from oat kernel (9).It is applied as a fat substitute in beef patt ies (10) and sausages (8).Inulin is another soluble dietary fi bre obtained mainly from chicory root and it is also known as a food prebiotic ingredient (11), applied in a variety of meat products (12,13).
Burgers and patt ies are common meat products, widely consumed in diff erent societies.Therefore, introduction of some changes in formulations in order to improve their health-related properties is of particular interest in the meat industry.
Mixture design methodology is a new approach that can be used to determine the functions of the ingredients in processed foods and provide valuable information about ingredient interactions by applying reduced numbers of experimental trials (14).As far as we know, the effects of incorporation of the combination of inulin and β-glucan on the physical and sensory properties of low--fat burgers containing vegetable oil have not been investigated previously.Hence, the objective of this research is to apply the D-optimal mixture design to determine the infl uence of the addition of inulin, β-glucan and breadcrumbs.Also, the eff ect of mixture interactions on the properties of low-fat beef burgers containing pre-emulsifi ed canola and olive oil blend was investigated in order to obtain the optimal mixture proportions to formulate healthy low-fat beef burgers.

Materials
Sixteen boneless cuts from heifer carcasses (brisket and fl ank) of mean age 17-24 months were obtained from Gooshtiran Co. (Tehran, Iran) at 48 h post mortem and stored in refrigerators (4 °C).All visible fat and connective tissue were removed.Proximate composition (in g per 100 g: moisture 71, protein 23, fat 5.5 and ash 0.9) of four portions of lean meat were analysed prior to burger manufacture.

Preparation of canola and olive oil blend
The oil-in-water emulsion (in g per 100 g: olive oil 2.4, canola oil 3.6, soya protein isolate 0.6 and water 4.8) was prepared with procedure described by Bloukas et al. (15).

Beef burger preparation
The formulation of low-fat beef burgers consisted of 48.5 % lean meat.It was minced using a grinder (model WWB 200; Laska, Traun, Austria) and the ingredients (in %: NaCl 1.2, black pepper 0.2, powdered cinnamon 0.05 and chopped onion 30) were added to the minced lean meat, then mixed for 5 min in a commercial mixer (model R6-02VB; Robot Coupe, Vincennes, France).
Then, the pre-emulsifi ed canola and olive oil blend (11.4 %) and selected mass fractions of powdered inulin, powdered β-glucan and bread crumbs were added to the mixture (Table 1).The mixture of ingredients was homogenized in the commercial mixer for 8 min to form burger paste.It was reground through the 6-mm plate, and then portions of 100 g of the burger mix were machine-shaped into a round mould (diameter 10 cm and height 1 cm).Beef burgers were then placed on stainless steel trays, wrapped with polyvinyl chloride plastic, and stored in boxes in commercial freezers (-18 °C) until the day of analysis.

Cooking procedure and cooking characteristics
Beef burgers were thawed at 5 °C for 12 h.The samples were cooked according to the American Meat Science Association methodology (16).A cooker (CFS FlowCook 600/6000; GEA, Bakel, The Netherlands) was used to heat the burgers to obtain an internal temperature of 71 °C.The cooking yield (η) of burgers was determined by measuring the mass of three patt ies for each formula before and aft er cooking using the following equation: The diameter (d) of three burgers of each formulation was measured at ambient temperature ((25±1) °C) before and aft er cooking.The following equation was used to calculate the diameter reduction: The texture properties of cooked beef burgers were evaluated using a texture analyser (M 350-10 CT; Testometric Co. Ltd., Rochdale, Lancashire, UK).Prior to analysis, beef burgers were thawed for 12 h at 5 °C.Texture profi le analysis test was performed using one portion of each beef burger (height 1 cm and diameter 2 cm) which was compressed to 50 % of its original height with a cylindrical probe of 3.6 cm in diameter and a cross-head speed of 2 mm/s.Force-deformation curves were applied to determine the textural measurements including hardness, cohesiveness and gumminess.Measurements were performed at ambient temperature ((25±1) °C).

Colour determination
Beef burgers were thawed for 12 h at 5 °C prior to colour determination.The colour values of burgers were determined with CIELab space values (L*=relative lightness, a*=relative redness and b*=relative yellowness) using a Chromo meter (CR-400; Konica Minolta Sensing, Inc., Osaka, Japan).Colourimeter was calibrated with the standard tile (L*=97.83,a*=-0.43,b*=+1.98).Colour properties were measured on three diff erent locations in each of the three examined samples.The calculated results were expressed with mean value of these measurements.

Sensory evaluation
Sensory analysis was done using a hedonic test as described by Mann and Whitney (17).Low-fat beef burgers used for sensory evaluation were cooked on a cooker (mentioned above), cut into four equally sized wedges, and one piece of each sample was placed on a white paper plate.Thirty untrained assessors were chosen among the students and staff of Food Science and Technology Department, Shahid Beheshti University of Medical Sciences, Tehran, Iran.The assessors were asked to sit in private booths with fl uorescent lighting at 25 °C and evaluate the overall acceptability of samples.Hedonic scores ranging from 1 to 9, representing a scale from extreme dislike to extreme like, respectively, were used for sensory evaluation.Mineral water and unsalted soda crackers were provided to cleanse the palate between the samples.
In order to evaluate the mouldability of burger patt ies, ten expert operators from one of the meat producing companies (Gooshtiran Co., Tehran, Iran) were invited to give their scores.A 5-point hedonic scale was used to evaluate this att ribute, with 1 being the least mouldable and 5 being the most mouldable.
Experimental design and optimisation D-optimal mixture design was used to investigate the eff ects of incorporation of inulin, β-glucan and breadcrumbs on cooking characteristics, textural, colour and sensory properties of low-fat beef burgers containing vegetable oils to determine the optimum mixture proportions in the formulation.Design-Expert ® soft ware v. 7.1.5(Stat --Ease Inc., Minneapolis, MN, USA) was applied to analyse the data and model the responses.Combined eff ects of breadcrumb (X 1 ), inulin (X 2 ), and β-glucan (X 3 ) content were evaluated using a six-level three-factor design.The mass fractions of these three ingredients varied from 0 to 8 g per 100 g of each burger.A total number of 14 experiments including 12 diff erent combinations of breadcrumbs, inulin and β-glucan content in two repetitions were performed according to the experimental design (Table 1).Numerical optimisation techniques of Design--Expert soft ware were used for the simultaneous optimisation of four responses including cooking yield, reduction in diameter, overall acceptability and mouldability of the burgers.Desired criteria for each variable and selected four responses were chosen.All the variables were kept in the range, while the responses were either maximised or minimised.Numerical optimisation was used to fi nd the best solution, with high desirability of the fi nal product.

Statistical data analysis
The following three equation models were fi tt ed to each of the responses (Y) with independent variables (X), where b is the regression coeffi cient calculated from the experimental data by multiple regressions.

Linear model: Y=b
The importance of all parts in the polynomial was evaluated statistically by computing the F-value at p=0.05 or 0.01.
All parametric tests were performed in triplicate for each experiment and the reported data were expressed as the mean value with standard deviation.The tests were analysed using the SPSS v. 17.0 for Windows (SPSS Inc., Chicago, IL, USA) and statistical signifi cance was calculated by Tukey's multiple range tests where p<0.05 or 0.01 was taken as the signifi cance.Sensory evaluation was analysed by the same soft ware according to the methods described by Mann and Whitney (17).Correlation analyses were conducted using the Pearson correlation model where p<0.05 or 0.01 was taken as signifi cant.

Cooking characteristics of low-fat beef burgers
The presence of β-glucan in low-fat beef burger formulations led to higher cooking yield values, whereas increasing the level of powdered inulin up to 8 % decreased the cooking yield (Table 2, formulation 1).As shown in Table 2, the highest cooking yield (76.1 %) and moisture retention (66.8 %) values were observed in formulation 11 containing 4 % β-glucan and 4 % breadcrumbs, while the lowest cooking yield (50.2 %) and moisture retention (36.2 %) values were found in the burgers containing 8 % inulin (Table 2).Signifi cant correlation between cooking yield and moisture retention was observed (R=0.99,p<0.01).
As it can be seen in Table 3, only the linear terms and the two-component interaction (between inulin and β-glu-can) had a highly signifi cant eff ect on cooking yield and moisture retention (p<0.01).Apparently, absolute value of breadcrumbs and β-glucan coeffi cients was greater than of inulin, which resulted in more signifi cantly positive effect of β-glucan and breadcrumbs in comparison with inulin on cooking yield and moisture retention.The contour plots of cooking yield and moisture retention show that the highest values of these properties were found in the β-glucan vertex and the lowest one appeared to be in the inulin vertex (Figs.1a and b).
Among the burgers, fat retention was higher in formulations 1 and 3, with 8 and 5.3 % inulin, respectively (Table 2), while it was lower in formulations 6 and 7, which contained only breadcrumbs or β-glucan, respectively.As shown in Table 3, the linear and non-linear terms had a signifi cant eff ect on fat retention (p<0.05).Fat  retention values were signifi cantly greater in the burgers made with inulin only, in those with the mixture of inulin and breadcrumbs, and those with the mixture of β-glucan and breadcrumbs than in other burgers (Fig. 1c).
A dimensional change is one of the most important alterations of burgers which can be aff ected by incorporation of new ingredients.In this research, the greatest reduction in the diameter was observed when 8 % inulin was added while the smallest change was detected when only β-glucan was added (Table 2).The adjusted regression coeffi cients showed that only the linear terms had a significant eff ect (p<0.01) on the diameter reduction of beef burgers (Table 3).In this way, coeffi cient value of inulin was greater than of breadcrumbs and β-glucan.The higher reduction in diameter values of burgers found in the vertex of inulin was confi rmed by contour plot (Fig. 1d).
As it has been mentioned, β-glucan, followed by breadcrumbs, had greater infl uence on cooking yield and moisture retention when compared with inulin.Losses of fat and moisture during cooking were minimised by the tight and porous networks of β-glucan which could entrap fat and water within the meat protein system.Likewise, Piñero et al. (10) demonstrated that β-glucan was effective in retaining moisture and fat and also increasing the cooking yield in low-fat beef patt ies because three-dimensional matrices are created within the meat protein system.The eff ect of the addition of the breadcrumbs on the properties of beef burgers can be att ributed to wheat starch.Swelled starch granules embedded in protein gel matrices form heat-induced texture during the cooking, which increased the water-binding properties of protein gel matrices (18).Zeidler et al. (19) found that incorporation of 10 % breadcrumbs into the Salisbury steaks reduced the yield losses to 6.6 % of its initial mass.Under the examined conditions, inulin, even at higher proportions, reduced the moisture retention and cooking yield probably because it does not form tight and porous gel structures in the burger system.Lower moisture retention and cooking yield of the formulation containing inulin could be att ributed to hydrolysation of the dissolved long-chain inulin molecules at temperatures above 80 °C into the smaller chain ones, which are unable to form a gel (20).This fi nding was in contrast to the results of previous research which found that incorporation of 6 % inulin in powdered form into a meat emulsion system decreased the cooking loss (21).As mentioned previously, inulin and β-glucan interaction had greater synergistic effect on moisture retention and cooking yield than the effect of interaction of other two components (the mixture of inulin and breadcrumb and/or the mixture of breadcrumbs and β-glucan).However, this eff ect was lower when only β-glucan was applied (p<0.05) when compared with when inulin and β-glucan mixture was incorporated in the formulation.A possible explanation is that the inulin chain might disturb the interaction of β-glucan with meat protein, and as a consequence, cause a signifi cant cooking loss.
All of the added ingredients (inulin, β-glucan and breadcrumbs) had a signifi cant eff ect on the changes in Fig. 1.Contour plots for the eff ect of the addition of breadcrumbs (X 1 ), inulin (X 2 ) and β-glucan (X 3 ) on cooking yield, moisture retention, fat retention and reduction in the diameter of the low-fat beef burgers containing canola and olive oil blend the diameter of beef burgers (p<0.05).Increasing the mass fraction of inulin in the formulations had the greatest infl uence on the beef burger diameter, followed by breadcrumbs and β-glucan.The smallest reduction in the diameter of samples containing β-glucan during cooking could be due to the stabilising eff ect of the β-glucan that holds meat particles connected to each other (10).In this study, a signifi cant correlation was found between the reduction in the diameter and moisture retention of beef burgers (R=-0.89,p<0.05).

Texture profi le of low-fat beef burger
The highest percentage of inulin (8 % in formulation 1) resulted in a harder structure; however, the lowest hardness value was observed in the burger containing 5.3 % inulin and 2.7 % β-glucan (formulation 3) (Table 4).The regression coeffi cient (Table 3) indicated that only the linear and quadratic terms were signifi cant (p<0.01).The effect of inulin, β-glucan and breadcrumbs on the texture was proportional to the corresponding coeffi cient (inulin> breadcrumbs>β-glucan) at the same percentage.Thereby, inulin was the most eff ective factor in increasing the hardness of the product.The predicted regression coeffi cient for the hardness showed that the interaction between inulin and β-glucan, breadcrumbs and β-glucan, and breadcrumbs and inulin had antagonistic eff ect on the hardness of the burgers.The contour plot of hardness indicated that the highest hardness values were in the inulin vertex and the lowest values were in the crossing of inulin and β-glucan interaction (Fig. 2a).
The highest cohesiveness was measured in the formulations containing only breadcrumbs (8 % in formulation 6) and the lowest value was recorded in the formulation containing equal mass fractions of inulin and β-glucan (formulation 4).The linear regression terms were highly signifi cant (p<0.01).In the predicted model, strong antagonistic eff ects of the two-component interaction on cohe-siveness were observed (p<0.01)(Table 3).The lowest cohesiveness values were observed in the crossing of β-glu can and inulin, whereas the highest values were found along the breadcrumb and/or inulin vertex (Fig. 2b).
The highest value of gumminess (23.4 N) was recorded in burgers containing only inulin (formulation 1).However, burgers containing 5.3 % inulin, 2.7 % β-glucan and 5.3 % β-glucan (formulations 2, 3 and 14, respectively) had lower gumminess values.As shown in Table 3, the linear terms and the two-component interactions had highly signifi cant eff ects on gumminess value (p<0.01).The absolute value of the coeffi cient of inulin was greater than of breadcrumbs and β-glucan, meaning that inulin had greater eff ect on gumminess value compared to two other ingredients.The contour plot confi rmed that the highest gumminess value was associated with inulin vertex (Fig. 2c).
The observed diff erences in textural profi les of 14 formulations can be att ributed to the diff erent impacts of the used hydrocolloids (inulin and/or β-glucan) or breadcrumbs as well as their interactions, mainly the possible interplay of added hydrocolloids with the starch content of breadcrumbs.Hydrocolloids can modify the gelatinisation of starch in the following ways: (i) decreasing the available water for swelling of starch and increasing the gelatinisation temperature, or (ii) molecular interaction with the starch granules to produce a more stable structure, which can result in an increase of the gelatinisation temperature (22).Suppressing the swelling of starch granules resulted in reduced amylose leaching, and consequently, disturbed network formation (22).Swelling of the starch granules surrounded by protein gel matrices resulted in the formation of strong heat-induced structure, and thereby, in the beef burger with a fi rm and more compact structure (23).Addition of β-glucan to the formulation containing breadcrumbs decreased the hardness, cohesiveness and gumminess of the burgers.This was probably due to reduced amount of available water for swelling of the starch granules, which led to an increase in gelatinisation temperature.
The addition of powdered inulin to the formulation caused higher moisture release during cooking, which changed the texture of burgers, i.e. increased hardness.Similar result was obtained by García et al. (24), who indicated that incorporating powdered inulin in the low-fat and full-fat mortadella increased textural hardness.Menegas et al. (12) showed that the addition of powdered inulin in reduced-fat sausages also resulted in increased hardness.
According to the regression coeffi cients in Table 3, it can be observed that the mixture of inulin and β-glucan in burger formulation decreases the hardness, cohesiveness and gumminess of the products.However, when the mass fraction of inulin in the mixture was higher than of β-glucan, these parameters increased.Possible explanation is related to the reduction in moisture retention of the burgers that amplifi es the compactness of the protein network (more fi rm and cohesive structure).In this study, the hardness of burger texture was signifi cantly correlated with the loss of moisture during cooking (R=0.87,p<0.01).

Diff erences in colour parameters of low-fat beef burgers
The highest L* value (lightness) was recorded in the low-fat beef burger containing only inulin (8 % in formu-lation 1), but the lowest lightness was observed in the formulation with 8 % β-glucan (formulation 7) (Table 4).The predicted regression analyses for the lightness showed that the linear terms were highly signifi cant (p<0.01) and the two-component interactions had a signifi cantly negative eff ect (p<0.05)(Table 3).The lowest L* value was in the β-glucan vertex, whereas the highest one appeared to be in the inulin vertex (Fig. 3a).Moreover, the highest a* value (redness) was obtained when the highest mass fraction of β-glucan was added to the burger mixture (formulation 7); the lowest a* value was determined in the sample with the highest mass fraction of inulin (formulation 1).For redness value (Table 3), the regression coeffi cients for the linear terms were highly signifi cant (p<0.01).In the contour plot, the lowest value was observed in the inulin vertex, whereas the highest value was recorded in the β-glucan vertex (Fig. 3b).The yellowness value (b*) was slightly higher in the burgers containing equal mass fractions of β-glucan and breadcrumbs or containing only 8 % β-glucan (formulations 11 and 7, respectively) and lower in the burgers containing equal mass fractions of inulin and breadcrumbs (formulation 9) (Table 4).According to the regression coeffi cient given in Table 3, the linear terms and the β-glucan and breadcrumbs interaction had highly signifi cant eff ect on yellowness values.The highest yellowness value in the predicted contour plot was obtained in the β-glucan vertex or when β-glucan was added to the formulation containing high mass fraction of breadcrumbs (Fig. 3c).Fig. 2. Contour plots for the eff ect of the addition of breadcrumbs (X 1 ), inulin (X 2 ) and β-glucan (X 3 ) on texture parameters of the low--fat beef burgers containing canola and olive oil blend Among the 14 formulations, the mass fraction and the type of fat remained constant, so the observed diff erences in colour values were probably due to the eff ect of added hydrocolloids and breadcrumbs individually and their interaction as well.In this study, higher moisture retention and higher cooking yield resulted in lower lightness and higher redness of beef burgers, which is in agreement with the fi ndings of García-García and Totosaus (25).Higher moisture release and higher cooking loss were signifi cantly correlated with lightness (R=0.916,p<0.01 and R=0.928, p<0.01, respectively).Since the addition of inulin to the burgers led to higher moisture release and lower cooking yield, the burgers containing inulin seemed to be lighter.However, the addition of β-glucan increased cooking yield and moisture retention and thus decreased the lightness.This fi nding was in contrast with the results obtained when β-glucan was added to the sausages, suggesting that β-glucan increased the lightness (26).Observed diff erences in yellowness values of the burgers were probably due to the diff erences in the natural colour of the added ingredients.

Sensory scores of low-fat beef burgers
Formulation 1 (containing 8 % inulin) received the highest overall acceptability score, while burgers with the highest mass fraction of β-glucan (formulation 7) received the lowest scores (Table 5).The regression coeffi cient val-ues of overall acceptability indicated that this response was aff ected signifi cantly by the linear terms (p<0.01)(Ta-  Values are mean±standard deviation.Mean values in the same column with the same lett er are not signifi cantly diff erent For the composition of the formulations see Table 1 ble 3).The lowest scores of overall acceptability in the contour plot were observed in the β-glucan vertex and a linear increase of overall acceptability scores was observed towards the edge of inulin and breadcrumbs interaction (Fig. 4a).The highest (4.9) and the lowest (1.2) scores for mouldability were given to formulations 6 and 7, respectively, by experienced panellists.The linear regression model was highly signifi cant for this response (p<0.01), and the highest regression coeffi cient value was att ributed to inulin, followed by breadcrumbs (Table 3).
In the contour plot, the lowest scores for mouldability were observed in the β-glucan vertex (Fig. 4b).Consequently, the addition of high mass fractions of inulin improved the mouldability and overall acceptance of the burgers compared with the addition of β-glucan.

Mixture proportion optimisation
The aim of the optimisation was to predict the optimal variables for some of the important responses which had a great infl uence on quality properties of burgers.The observed and predicted values of the responses along with the mixture proportions are shown in Table 6.
Our aim was to maximise cooking yield, mouldability and overall acceptability, and minimise the reduction in diameter of the burgers, while maintaining the mass fractions of inulin, breadcrumbs and β-glucan within range.By using the given criteria, two mixtures were suggested (mixture 1: 2.7 % breadcrumbs, 3.1 % inulin and 2.2 % β-glucan; and mixture 2: 5.8 % inulin and 2.2 % β-glucan) (Table 6).Mixture 1 was selected as the best because it received higher (71.6 %) desirability score (Table 6).The optimum formulation could provide valuable nutritional and technological properties in comparison with traditional beef burgers.Furthermore, introducing the changes in the formulation did not have negative eff ects on sensory properties of the product.

Conclusion
D-optimal mixture design approach was used to optimise the formulation of beef burgers.According to the experimental results and the map of contour plots, addition of inulin to the formulation had a predominant eff ect on cooking characteristics, textural and colour parameters of cooked burgers.Addition of inulin to the formulation of low-fat beef burgers decreased cooking yield and moisture retention; furthermore, it increased the lightness, mouldability, overall acceptance of the burgers and the textural parameters, while adding β-glucan had inverse eff ects.In this study, the interaction between inulin and β-glucan improved the cooking characteristics of the product without having a signifi cantly negative eff ect on the colour and sensory att ributes of the burgers.Therefore, according to the obtained results, the mixture of inulin and β-glucan can be used in the formulation of low-fat beef burgers.This mixture not only off sets some negative eff ects att ributed to the fat level reduction and total replacement of beef fat by vegetable oils, but also improves some cooking properties and texture of the products along with enhancing nutritional characteristics.

Fig. 4 .
Fig. 4. Contour plots for the eff ect of the addition of breadcrumbs (X 1 ), inulin (X 2 ) and β-glucan (X 3 ) on overall acceptability and mouldability of the low-fat beef burgers containing canola and olive oil blend

Table 1 .
D-optimal mixture design used to study the eff ects of the addition of breadcrumbs, inulin and β-glucan (at total mass

Table 2 .
Experimental results for cooking characteristics of each burger formulation aValues are presented as mean±standard deviation of three repetitions.Mean values in the same column with the same lett er are not signifi cantly diff erent For the composition of formulations see Table1

Table 4 .
Experimental results for texture and colour parameters of each burger formulation abcdValues are mean±standard deviation of three repetitions.Mean values in the same column with the same lett er are not signifi cantly diff erent.L*=lightness, a*=redness, b*=yellowness For the composition of formulations see Table1

Table 5 .
Scores of overall acceptability and mouldability of low--fat beef burgers

Table 6 .
Optimum levels of independent variables along with predicted and observed values of the responses 1 =breadcrumbs, X 2 =inulin, X 3 =β-glucan OV=observed value, PV=predicted value edge Gooshtiran Co., Tehran, Iran, for their help in manufacturing burgers. X