Optimization of synbiotic yogurts with yacon pulp ( Smallanthus sonchifolius ) and assessment of the viability of lactic acid bacteria

The market of food with functional properties in Brazil invoice on average, US$ 10 billion a year, and worldwide, the segment is expected to grow by 38% by 2017, according to the Brazilian Asociación of Diet Food Industries and Special purposes (Asociação Brasileira das Indústrias de Alimentos Dietéticos e para Fins Especiais, 2016). The dairy industry has been outstanding in this respect with the largest number of functional products by addition of probiotics and prebiotics in foods such as yogurt and fermented milk generally (Wendling & Weschenfelder, 2013; Santos et al., 2011; Sanchez et al., 2009).


Introduction
The market of food with functional properties in Brazil invoice on average, US$ 10 billion a year, and worldwide, the segment is expected to grow by 38% by 2017, according to the Brazilian Asociación of Diet Food Industries and Special purposes (Asociação Brasileira das Indústrias de Alimentos Dietéticos e para Fins Especiais, 2016).The dairy industry has been outstanding in this respect with the largest number of functional products by addition of probiotics and prebiotics in foods such as yogurt and fermented milk generally (Wendling & Weschenfelder, 2013;Santos et al., 2011;Sanchez et al., 2009).
Probiotic and synbiotic dairy products are leaders in the functional foods market and are considered research priorities in many countries.The valorization of functional foods worldwide is fostering innovations in food products and stimulating the consumption of foods with important nutritional and therapeutic value (Kumar et al., 2015;Saccaro et al., 2009).
Among the dairy products, yogurt/fermented milks have been the subject of several studies all over the world and different benefits for human health have been reported after their ingestion (Wang et al., 2012).The regular consumption of probiotics foods can provide the equilibrium of the intestinal microbiota and health benefits (Saad et al., 2013;Vitali et al., 2012).It has been demonstrated that consumption of probiotic can enhance the immune system (Lollo et al., 2013(Lollo et al., , 2015)).
To improve the functionality of the food matrix and facilitate entry of functional foods market, many dairy products are being developed with non-digestible oligosaccharides such as inulin and fructo-oligosaccharides (FDS) as prebiotic ingredients.As such, yacon (Smallanthus sonchifolius) is considered to be the best natural source of FDS and is added to many products to promote healthy intestinal microbiota as these oligosaccharides are selectively fermented in the colon by bifidobacteria and lactobacilli and can therefore provide numerous local and systemic benefits (Dliveira et al., 2016;Sousa et al, 2015;Delgado et al., 2013;Charalampopoulos & Rastall, 2012;Roberfroid, 2007).The tuberous root of yacon is perennial, thrives in many Andean countries of South America, is generally eaten raw and its colour varies from light to dark yellow (Shi et al., 2015;Velez et al., 2013;Djansivu et al., 2011).
Many studies have been made with prebiotics, Cruz et al. (2013a) using oligofructose as a partial replacement of sucrose, developed a plain yogurt and obtained a low calorie product provide an increase in the nutritional benefits without compromising their technological properties.Batista et al. (2015) developed probiotic yogurts with quality parameters positives (probiotic bacteria count, physical-chemical and metabolic activities) with interesting technological option for small and medium-size dairy enterprises to enter to the market of functional dairy foods.Cruz et al. (2013b) results, suggest that the use of packaging systems with different oxygen permeability rates coupled with the addition of glucose oxidase presented an interesting technological option to minimize the oxidative stress in probiotic yogurts.Balthazar et al. (2016) added inulin to sheep milk yogurt smoothies (SMY) and concluded that there was a decreased in post-acidification which aid in bacteria survival and the viability in the food matrix.Morais et al. (2014) developed a prebiotic chocolate dairy dessert adding different concentrations of prebiotics (inulin and FDS) and observed that sensory properties not affected by adding 7.5% of prebiotics.
Dverall, the addition of both prebiotics and probiotics to yogurts has increased their sensory acceptability over products containing only probiotics and conventional products (Allgeyer et al., 2010), reinforcing the idea that the properties of functional foods containing prebiotics and probiotics should be similar to those of the original product (Cruz et al., 2010).However, the probiotic properties will be preserved only if these products maintain a good sensory acceptability and if microorganisms remain viable during shelf life (Komatsu et al., 2008).Thus, the assessment of bacterial viability in a probiotic product is critical for quality evaluation.In this respect, previous studies reported that some commercial products exhibited unacceptably low levels of viable bacteria (Al-Dtaibi, 2009;Moreno et al., 2006;Carr & Ibrahim, 2006).
The amount of ingredients such as sugar and milk powder can influence the viability of probiotic bacteria and the acceptability of yogurts by directly altering their flavour and texture (Dave & Shah, 1997).Studies that purport to investigate the viability of probiotic bacteria in fermented skim milk produced with different levels of milk powder and sugar can be useful for small dairy industries that are interested in producing flavoured fermented milks without decreasing the viability of probiotic micro-organisms during its shelf life (Maganha et al., 2014).
Therefore, sensory assessments must be conducted with consumers to define the necessary adjustments in the formulation (Ares et al., 2009;Minim, 2013).In this respect, the response surface methodology (RSM) is an effective statistical technique for the establishment of models and processes in which the response of interest, such as sensory acceptability and intent to purchase a product, is influenced by several other variables simultaneously (such as product ingredients), leading to a more efficient and accurate optimization process by reducing the number of experimental studies and product development time and overall research costs (Myers & Montgomery, 2002).Importantly, RSM efficiently optimizes ingredients and other variables (Ghasemlou et al., 2013;Shen et al., 2011;Granato et al., 2009;Acosta et al., 2008).
Here we aimed to study the effects of sugar, yacon pulp and skimmed milk powder (independent variables) in the aroma, flavour, texture, appearance, overall impression and intent to purchase (dependent variables) on synbiotic yogurts to optimize the formulation using RSM.The most accepted yogurt formulations were subjected to viability tests by counting both the traditional bacteria and the probiotic bacteria added.

Preparation of the yacon pulp
Yacon roots were purchased from Ceasa Supply Center of Pernambuco.The other ingredients were obtained from the local market in Recife, Pernambuco State, Brazil.
Roots were washed and sanitized in 200 ppm chlorine solution for 15 min, peeled, cut into 2-cm cubes and soaked in ascorbic acid solution (0.5 g/100 g) for 5 min.Subsequently, these cubes were bleached at 100 °C for 5 min, cooled on ice for 2 min and ground in a food processor.The pulp was concentrated up to 25 °Bx.

Preparation of synbiotic yogurts
Yogurts were prepared with traditional lactic cultures (Streptococcus salivarius thermophilus and Lactobacillus delbrueckii ssp bulgaricus, Danisco YD-MIX 499) and probiotics (L.paracasei ssp paracasei LBC 81, Danisco), which were individually and aseptically transferred to 1 L of skim milk and stored at −18 °C.For the inoculation step, cultures were thawed to 37 °C and 2 mL of each culture was added to 1 L of milk for each formulation.
For each experiment, different concentrations of sugar, skimmed milk powder, yacon pulp and 0.3% of clear unflavoured powdered gelatin were added to 1 L of UHT whole milk.The independent variables used were as follows: percentage of yacon pulp, X 1 (5 to 12%); percentage of sugar, X 2 (4 to 9%) and percentage of skimmed milk powder, X 3 (1 to 6%).These ingredients were homogenized and pasteurized at 95 °C for 2 min, and the mixtures were cooled to 37 °C and maintained in an incubator at 37 °C until the pH reached 4.4, after which they were stored at 10 °C.

Sensory assessment
The sensory panel was recruited from 50 potential panellists who were firstly subjected to a preliminary screening by a questionnaire, were then further screened for their ability to discriminate between basic tastes based on threshold testing as outlined by the International Drganization for Standardization (2007).All participants were familiar with yogurt taste and had previous experience with sensory evaluation.35 trained adult from the Center for Health Sciences at the Federal University of Pernambuco (23 women and 12 men, 26 to 38 years old) comprising teachers, students and staff, were selected.Panellists were asked to evaluate the aroma, flavour, texture, appearance, overall impression and intent to purchase.Therefore, we used a structured hedonic scale with nine points to grade each attribute, in which the value of 1 represented 'extreme dislike' and that of 9 represented 'extreme like' .For the intent to purchase attribute, a 5-point scale was used, in which the value of 1 represented 'definitely would not buy' and the value of 5 represented 'definitely would buy' (Pandey & Mishra, 2015).Approximately 40 g of yogurt were given to the participants in white plastic cups coded with three digits.Panellists used water to rinse their mouth before tasting different samples.Assessments were performed in individual booths and were recorded on a sensory evaluation sheet (Minim, 2013;Pimentel et al., 2013;Meilgaard et al., 2007).
A balanced incomplete block design type III was applied in this study with the following parameters (Hinkelmann & Kempthorne, 1994;Cochran & Cox, 1992;Corrêa et al., 2008).

Response Surface Methodology (RSM)
For the formulation of synbiotic yogurts, a 2 3 factorial design was constructed to investigate the effects of three independent variables on product acceptability and intent to purchase: concentration of yacon pulp (X 1 ), percentage of sugar (X 2 ) and percentage of skimmed milk powder (X 3 ).These variables were analysed in two equidistant levels of variation (coded as −1 and +1), three centre points (0) and six axial points (± α, 0), (0, ± α), totalling 17 experiments (Table 1) (Rodrigues & Iemma, 2009;Drake et al., 2010).The dependent variables (responses) were aroma, flavour, texture, appearance, overall impression and intent to purchase, and served as quality parameters for the formulations.The data obtained were fitted to the following polynomial equation (Equation 1): where βn are the regression coefficients; y is the response in question (sensory attributes and intent to purchase) and P, A, and L are coded variables (percentages of yacon pulp, sugar and milk powder, respectively).

Viability of lactic acid bacteria during storage
The formulations with the highest scores in sensory acceptability and intent to purchase were subjected to a viability test during the 21-day storage period at 10 °C.For this purpose, a 1-mL aliquot of each sample was transferred to a tube containing 9 mL of 0.1% sterile peptone water.This initial dilution was used to make decimal serial dilutions of the yogurts up to the dilution of 10 7 .After incubation for each culture medium was complete, cell count was performed in Petri dishes.
The number of viable cells was determined using the pour plate method with 1-mL aliquots.S. thermophilus was counted on the M17 agar medium with the addition of 5% lactose solution at 10%, followed by aerobic incubation at 37 °C for 48 h.For counting L. delbrueckii ssp.bulgaricus, MRS agar was used after adjusting the pH to 5.4 with glacial acetic acid, followed by anaerobic incubation at 37 °C for 72 h in GasPak™ jars containing an anaerobic generator (International Dairy Federation, 1988).
Direct count of L. paracasei ssp.paracasei LBC 81 was performed using MRS vancomycin agar (Tharmaraj & Shah, 2003).Approximately 2 mL of a vancomycin solution (0.05 g/100 mL), previously sterilized by filtration through 0.45-μm membranes (Millipore), was added to 1 L of MRS agar to obtain a final concentration of 1 mg/L.Plates were incubated anaerobically at 37 °C for 72 h.

Ethical considerations
This research was approved by the research ethics committee of the Federal University of Pernambuco according to the standards and regulatory guidelines for research involving humans through Resolution CNS 196/96, CAAE -0345.0.172.000-11.Participation in the study was voluntary and participants were informed about the research and methodological details by reading and signing the Term of Free and Informed Consent established by the National Health Council.

Sensory assessment
Table 2 shows results of the 2 3 factorial planning for the dependent variables, sensory analysis and intent to purchase.Table 3 indicate that the yacon pulp had a statistically significant negative influence (p < 0.05) in the linear term in all sensory attributes and intent to purchase, with greater influence on flavour, appearance, overall impression and intent to purchase, i.e. smaller percentages of yacon contributed to the achievement of higher scores (Table 2).
The linear term for the sugar concentration had a positive and larger influence on flavour, texture, overall impression and intent to purchase, i.e. higher percentages of sugar contributed to the achievement of the highest scores for these quality parameters.
With regard to the milk powder, this linear term had a greater influence only on texture and intent to purchase (Table 3).We should also emphasize that the interaction between the yacon pulp and skimmed milk had the greatest influence on these attributes.
Table 2 indicates that the highest percentages of sugar and the lowest percentages of yacon pulp contributed to the higher scores exhibited by formulations 3, 7 and 12, differently from formulations 1 and 5, which had lower percentages of sugar and yacon pulp.
Formulations 3, 7 and 12 exhibited scores close to 7 for overall impression, meaning that participants liked the formulation moderately and a score close to 4 for intent of purchase, indicating that participants would probably buy the product (Table 2).
Models were calculated for each of the dependent variables and only statistically different parameters were included (p < 0.05) (Table 4).The models that were adjusted for aroma, flavour, texture, overall impression and intent to purchase showed coefficients of determination and coefficient of appearance to be greater than 0.99 and equal to 0.89, respectively.
The proximity of the coefficient of determination of the equation to 1.0 demonstrated the validity of the regression equation and hence the absence of non-significant fit for each model (p > 0.05).Therefore, the coefficients of determination for the predicted model for aroma, flavour, texture, appearance, overall impression and intent to purchase were valid, considering that response variables were hedonic sensory measures, which often exhibit considerable variations.
Figure 1A-H shows the construction of response surfaces according to each model.Figure 1A shows that the surface created by the predictive model for flavour acceptability enabled better determination of formulation optimization using a lower percentage of yacon pulp and a higher percentage of sugar.The highest scores for flavour were obtained with 7 to 9% of yacon pulp and 8 to 10% of sugar.With regard to the relationship between yacon pulp and sugar, the same behaviour could be observed in relation to texture, overall impression and intent to purchase, as shown in Figure 1B, F, H, respectively.These results corroborate those found in the estimate of the effects (Table 3).
Figure 1C shows that the highest scores for texture were obtained with a higher percentage of milk powder and the lowest scores were obtained for appearance in relation to sugar concentration (Figure 1D).
Figure 1E shows that the smallest percentages of yacon pulp and the largest percentages of milk powder exhibited the greatest influence on product appearance, corroborating the negative effect of the interaction between yacon pulp and milk powder (Table 3).

Culture viability during storage
Table 5 shows the direct counts for the traditional lactic acid bacteria S. salivarius ssp.thermophilus and L. delbrueckii ssp.bulgaricus and for the probiotic culture during storage at 10 °C.Formulations 3, 7 and 12 exhibited the highest scores of acceptability and intent to purchase and were designated A, B and C, respectively.In general, formulation B permitted a lower bacterial growth.At the 21 st day of storage, S. thermophilus predominated over the other microorganisms.Moreover, probiotic bacteria remained viable throughout the shelf-life period.

Sensory assessment
The highest percentage of sugar in the yogurt formulations positively influenced the flavour, texture, overall impression and intent to purchase.In general, yogurts containing fruit pulp and higher contents of sugar are preferred by the Brazilian consumers.Although sugar was added to products, FDS may have contributed only slightly to the sweetness of the formulations, as it exhibits only 30 to 35% of the sweetness of sucrose (Coussement, 1999;Franck, 2006).Notably, flavour is one of the most important factors influencing the decision to purchase foods (Drewnowski & Darmon, 2005;Glanz et al., 1998).
The little influence of the yacon pulp on the aroma may be related to the high moisture content of this root.In Quechua, both yacu and unu indicate water, whereas yakku indicates watery or tasteless, hence the neutral and smooth aroma of yacon (Delgadoet al., 2013;Moscatto et al., 2006).
Sugar most influenced the texture of yogurt.We also observed that the texture increased with a higher percentage of milk powder and a lower percentage of yacon pulp.This result contradicts the literature, which has reported that the presence of FDS is responsible for the increase in food texture (Saad et al., 2013).Dn the other hand, Karimi et al (2015) mentions that when FDS added to food in low concentrations, the rheological properties and the sensory quality of the product will not be affected strongly due to FDS neutral or slightly sweet taste and its limited effect on viscosity.
Because of the Maillard reaction, skimmed milk powder was the ingredient that most influenced the appearance of the formulations.We also observed that the fructo-oligosaccharides present in yacon are non-reducing sugars and hence are not susceptible to the Maillard reaction (Niness, 1999).
This study corroborates the data found in previous studies with regard to the high coefficients of determination.In an assessment of a low-calorie gelatin made of tropical fruits, by varying the percentage of sweetener, pectin with a low degree of methoxylation, and calcium, the response variables yielded significant adjusted models (p = 0.014) and no significant lack of fit (p = 0.253) (Acosta et al., 2008).The acceptability of low-fat, prebiotic dairy beverages was evaluated by 50 customers and with a 9-point hedonic scale using short-and long-chain inulins.The model indicated a coefficient of determination of 0.87 for the formulation prepared with short-chain inulin and a coefficient of 0.86 for long-chain inulin, and neither of them showed lack of fit, with values of p = 0.217 and p = 0.023, respectively (Villegas et al., 2010).

Viability
Total counts of the lactic bacterial strains S. thermophilus and L. delbrueckii ssp.bulgaricus, which are common in yogurt, were above the required minimum of 10 7 viable microorganisms per gram (Food and Agriculture Drganization of the United Nations, 2003).This density meets the standards established by the Brazilian legislation, and according to the Standards of Identity and Quality of Fermented Milk (through Normative Instruction 46 from 10/23/2007), must be at least 10 7 UFC/g in the final product throughout the validity period for the fermented milk to be considered a yogurt (Brasil, 2007).
The larger percentages of sugar and milk powder used in formulation B may have led to a decrease in the amount of water and hence decreased the ability of microorganisms to multiply.The solid content and sugar concentration (i.e.osmotic pressure) have been shown to directly influence the population of lactic acid bacteria (Shah & Ravela, 2000;Vinderola et al., 2000).Therefore, the results of this study corroborate previous reports.Although our formulations had 8 or 9% of sugar, this percentage did not hinder the growth of lactic acid bacteria to the point of yielding a low bacterial count.
In beverages fermented with S. thermophilus, L. delbrueckii ssp.bulgaricus, Bifidobacterium and L. acidophilus, S. thermophilus has been reported to predominate over the other microorganisms, similar to what we observed in the present study (Thamer & Penna, 2005).
The total counts of L. paracasei ssp.paracasei LBC 81 were above the threshold for probiotic microorganisms added to yogurts (Food and Agriculture Drganization of the United Nations, 2003).Also, their survival until the 21 st day of storage at 10 °C suggests that our yogurt formulations provided a good food matrix for the growth and viability of this microorganism.The legislation of many countries considers that the bacterial population must be viable during storage, but the minimum levels established for an effective quantification may vary from 10 6 to 10 8 UFC/g, according to local regulations.
The Brazilian legislation has established that the minimum amount of viable probiotic bacteria in yogurts should be between 6 and 7 log CFU/100g/day (Brasil, 1999(Brasil, , 2003)).The three yogurt formulations used in this study were considered 'probiotic' during a shelf life period of 21 days, confirming what has been reported on the positive interaction between L. paracasei ssp.paracasei LBC co-cultured with S. thermophilus and L. delbrueckii spp.bulgaricus.This co-culture application also has contributed to the production of a final product with satisfactory aroma, flavour, texture and acidity characteristics (Kristo et al., 2003;Saarela et al., 2000).It has also been reported that the viability of the probiotic strain L. paracasei ssp.paracasei B117 co-cultured with S. thermophilus Y4.10 and L. delbrueckii spp.bulgaricus Y6.15 in fermented milk was greater than 6 log CFU/g during storage at 4 °C for 21 days (Kristo et al., 2003).
Previous data have indicated that the use of growth factors as prebiotics (such as FDS in yacon pulp) and milk proteins are very important for the maintenance of bacterial viability because these growth factors affect the kinetics of acidification (Nogueira, 2009;Dliveira et al., 2002).Corroborating our results, the addition of milk proteins has stimulated the multiplication of L. rhamnosus and L. paracasei (Saxelin et al., 1999).In a previous study where prebiotics, inulin and polydextrose were added to yogurts made with banana puree and L. paracasei, the number of probiotic bacteria remained high at 8.86 log CFU/g during the 21-day storage period (Srisuvor et al., 2013).5 Conclusions The percentage of yacon pulp had a pronounced negative influence on flavour, texture, appearance, overall impression and intent to purchase.The percentage of sugar had a greater positive influence on flavour, texture, overall impression and intent to purchase.Model predictions adjusted to the variables aroma, flavour, texture, appearance, overall impression and intent to purchase exhibited good predictive ability and hence could be used as tools for process control.In the yogurt formulations studied, traditional bacterial strains and L. paracasei ssp.paracasei LBC 81 remained viable throughout the storage period.
t = number of treatments = 15 k = number of units per block = 3 r = number of repetitions for each sample = 7 b = number of blocks = 35 (number of assessors) λ = the number of times that each treatment appeared in the same block together with each of the other two treatments = 1 This design resulted repeated in 70 different trials (N = 70) or responses needed to obtain accurate estimates.Each panellist received three samples of three different formulations, which were presented on different days.
(0.12) abc Means followed by the same lowercase letters vertically along the days of storage in the same yogurt formulation indicate non-significant differences (p > 0.05) by the Duncan test.ABC Means followed by capital letters vertically between the yogurt formulations on the same day of storage indicate non-significant differences (p > 0.05) by the Duncan test.* Mean and standard deviation are shown in parentheses.UNC: Uncountable; CFU/g: Colony Forming Units/g.

Table 1 .
Experimental design for the preparation of symbiotic yogurts.

Table 2 .
Results of the 2 3 factorial planning for the dependent variables 'sensory analysis' and 'intent to purchase' .

Table 3 .
Estimate of the effects on the sensory attributes (aroma, flavour, texture, appearance and overall impression) and intent to purchase.

Table 4 .
Estimated models for the sensory attributes and the intent to purchase.
P: yacon pulp and S: sugar; M: skimmed milk powder; R 2 : coefficient of determination.

Table 5 .
Viability of bacteria (log CFU/g) used in yogurt fermentation during refrigerated storage at 10 °C for 21 days.