Influence of the Different Oils Used in Dough ...

Influence of the Different Oils Used in Dough Formulation on the Lipid Fraction of Taralli Abstract: An experimental investigation was carried out to evaluate the quality of taralli as a function of the type of oil

used in their dough formulation. To this end, 4 types of oil (extra virgin olive oil, olive oil, olive-pomace oil, and refined palm oil) were utilized to prepare taralli to investigate on the lipid fraction degradation and evaluate the taralli acceptability by consumers. The data obtained pointed out that taralli manufactured with extra virgin olive oil were significantly more appreciated than those made with refined oils due to their visual appearance and odor. Moreover, with respect to the other kinds of oil, the use of extra virgin olive oil led to significantly lower values (P < 0.05) of specific absorption at 232 and 270 nm (K232 and K270 , respectively) and of triacylglycerol oligopolymers. It also proved to present a much lower content in oxidized triacylglycerols and diacylglycerols than olive-pomace oil and refined palm oil, respectively. Furthermore, trans fatty acid isomers were absent in taralli made with extra virgin olive oil but were constantly present in those produced with refined oils. Keywords: HPSEC analysis, lipid fraction, oxidative degradation, taralli

Introduction Lipids play a primary role from a nutritional point of view and exert a fundamental action in many biological processes (Fernie 2003; Li and others 2003) since they are constituents of the cell membrane, carriers of fat soluble vitamins, and a source of energy and essential nutrients. Besides the nutritional aspects, lipids are also important in food processing. In particular, they influence the rheological properties, the sensory profile, and the palatability as well as the shelf life of bakery products (Pomeranz 1971; Larsson 1994). For this reason, a wide choice of edible fats and oils (margarines, shortenings, vegetable fats and oils, hydrogenated fats) is usually taken into consideration. They have different properties and are suitable to produce the various kinds of bakery products. The choice of the type of lipid to be used is often driven by technological and economic reasons and does not consider any nutritional aspects. Previous research carried out by Caponio and others (2009) on taralli prepared using palm oil highlighted the influence of the production process and of the storage time on the quality of the lipid fraction of taralli. In particular, the lipid fraction of the finished product was poorer in polyunsaturated fatty acids and richer in polymeric oxidation products when compared with the raw materials used, while storage resulted in a further increase in the fat oxidation products. It is well-known that high amounts of polymeric products, deriving from the oxidation of triacylglycerols, may have adverse effects on the consumers’ health (Billek 2000; Saguy and Dana 2003) and exert a pro-oxidant activity (Frankel and others 1988; Gomes and others 2008). Hence, they might negatively affect the shelf life of the product. Furthermore, the investigation showed the presence of trans isomers of unsaturated

fatty acids, that have negative effects on the consumers’ health (Koletzko and Decsi 1997; Morrison and others 2008), and whose content was significantly affected neither by the production process nor by the subsequent storage of taralli, but was solely depending on the quality of the oil used in the formulation. Hence, it is necessary to replace the fats frequently used in the recipes of these products with oils richer in antioxidants and whose composition accords to the indications of nutritionists. This study was therefore aimed at assessing the quality of taralli made with different types of oil, studying the degradation of the lipid fraction and evaluating the consumers’ acceptability.

Materials and Methods

Sampling The taralli making process consisted in mixing flour (1 kg), water (0.4 l), oil (0.2 kg), salt (0.02 kg), and spices (seeds of Foeniculum vulgare, 0.005 kg) with a diving arm kneader (Polin, Verona, Italy) for 20 min at about 20 ◦ C. Then, the dough was left to rest for 45 min. After this time, the dough was put into the so-called “tarallatrice,” a machine that cuts and rolls strips of dough and makes them enter mobile tubes, which by moving give the dough the famous taralli shape, that is, round with a hole in the middle, of about 3 cm in diameter and 8 mm in thickness. The resulting ring-shaped pieces of dough were then rapidly dipped into boiling water for a short time (about 2 min) and placed onto metal grid trays. The trays were then left at room temperature for 30 min to allow the surface to dry, before putting them into a tunnel air oven (Polin, Verona, Italy) at 230 ◦ C for about 20 min to bake and dry the product until a homogeneous crispy texture was achieved. Two different trials were made to compare the effect of the most common oils used in taralli making. In particular, for each trial, 4 different types of oil were used: olive oil (OO; refined olive MS 20101166 Submitted 10/15/2010, Accepted 1/25/2011. Authors are with oil blended with virgin olive other than lampante oil, in undeDipt. di Biologia e Chimica Agro-forestale e Ambientale (DIBCA), Univ. degli Studi, Via Amendola, 165/a, I-70126 Bari, Italy. Direct inquiries to author Caponio fined quantity), olive-pomace oil (OPO; refined olive-pomace oil added with virgin olive other than lampante oil, in undefined quan(E-mail: [email protected]). tity), refined palm oil (RPO)—all generally used in the production R 2011 Institute of Food Technologists doi: 10.1111/j.1750-3841.2011.02113.x

C

Further reproduction without permission is prohibited

Vol. 76, Nr. 4, 2011 r Journal of Food Science C549

C: Food Chemistry

Francesco Caponio, Mariagrazia Giarnetti, Carmine Summo, and Tommaso Gomes

Quality of the lipid fraction of Taralli. . .

C: Food Chemistry

process of taralli—and extra virgin olive oil (EVOO), that is seldom used for those kind of productions. The latter is an oil “obtained from the fruit of the olive solely by mechanical or other physical means that cause no deterioration of the oil” (Official Journal of the European Communities 2001). In the 2 trials, for each type of oil different lots were used to evaluate their effect on the taralli making, thus avoiding that differences observed could be attributed to the peculiar composition of the oil used. By contrast, the flour used was always from the same lot. All samples were analyzed within 1 d from baking.

mL/min. The HPSEC system consisted of a series 200 pump (Perkin–Elmer, Norwalk, Conn., U.S.A.) with Rheodyne injector, a 50-μL loop, a PL-gel guard column (Perkin–Elmer, Beaconsfield, U.K.), 5 cm length × 7.5 mm i.d., and a series of 2 PL-gel columns (Perkin–Elmer) 30 cm length × 7.5 mm i.d. each. The columns were packed with highly cross-linked styrenedivinylbenzene copolymer with particle of 5 μm and a pore di˚ The refractive index detector was a series 200A ameter of 500 A. (Perkin-Elmer). Peaks identifications and quantification were carried out as described elsewhere (Gomes and Caponio 1999).

Analyses of flours Protein (total nitrogen × 5.7), total ashes and moisture contents were determined according to the AACC approved methods nr 46–11A, 08–01, and 44–15A, respectively (AACC 2000). Dry gluten content and gluten index were assessed by means of the Glutomatic System (Perten Instruments, Hamburg, Germany) as in the AACC method nr 38–11 (AACC 2000).

Routine analyses The determination of the free fatty acids (FFA), peroxide value (PV), and spectrophotometric constants was carried out according to the Official Journal of the European Communities (1991).

Fat extraction The lipid fraction from fresh taralli samples and flour was cold extracted according to the method proposed by Folch and others (1957). In particular, lipids were extracted by homogenization of the samples with a methanol/chloroform mixture (1/2, v/v), followed by filtration and washing. Total fat was determined by acid hydrolysis and subsequent extraction of hydrolyzed lipid material, according to the AACC method nr 30–10 (AACC 2000). Preliminary trials indicated that the amount of fat recovered via solvent extraction accounted for about 95% of total fat obtained by acid hydrolysis. However, the Folch extraction was preferred because it does not require any heating steps thus permitting not to alter the fat quality to be evaluated by the following analyses. Determination of fatty acid composition and trans fatty acid isomers The determination was carried out by gas-chromatographic analysis of fatty acid methyl esters, according to the official methods (Official Journal of the European Communities 1991, 1992). In particular, the extracted fat was treated with methanol/hydrochloric acid (98/2, v/v), in a sealed phial, at 100 ◦ C for 40 min. After cooling, the phial was opened and 2 mL of distilled water and 1 ml of hexane were added. Then, 1 μL of the hexane fraction was injected. The gas-chromatographic system was composed by a Fisons (Milan, Italy) HRGC mega 2 series, equipped with a FID detector and a SP 2340 fused silica capillary column (Supelco, Bellefonte, Pa., U.S.A.), 60 m length × 0.25 mm i.d. and 0.20 μm film thickness. The temperature of the split injector was 230 ◦ C, with a splitting ratio of 1 : 100; the detector temperature was 280 ◦ C. The oven temperature was programmed as it follows: at the very beginning, the temperature was set at 140 ◦ C for 3 min, then gradually raised to 230 ◦ C (at 4◦ C/min) and held constant for 5 min. Helium was used as carrier gas at a flow of 1.5 mL/min. Several analyses carried out on the same sample produced a repeatability coefficient with a percent variation (CV %) of about 5%. Determination of polar compounds Polar compounds (PC) of the extracted lipid fraction were separated by silica gel column chromatography according to the AOAC method nr 982.27 (2003). After being recovered in THF, they were analyzed by means of high performance size-exclusion chromatography (HPSEC) using THF as eluant at a flow rate of 1 C550 Journal of Food Science r Vol. 76, Nr. 4, 2011

Consumer test Preference ranking has become an indispensable tool for food technologists to compare the taste qualities of similar food products. Participants (n = 80) were selected among people who regularly consume taralli (that is, 1 to 3 times a week). The sensory evaluation was done within 4 h from baking. A 3-digit random code was assigned to the 4 specialties of taralli that were served at room temperature in white polyethylene trays. For each sample, 20 g of taralli were served. Participants also received an evaluation form. After trying several different palate cleansers to avoid carryover effects and adaptation to sensory stimuli, mineral water, and 1 min break between samples were chosen. Before starting the evaluation of products, the investigator presented and read aloud a text to all participants to set up the assessment context. Their task was to judge the visual appearance, odor, taste, friability and, eventually, the overall acceptability (see Table 2) of four kinds of taralli and arrange the products in decreasing order of preference. Sensory evaluations took place in a conference room where temporary partitions were erected to create isolated tasting booths to separate testers during analysis, in compliance with the Standard nr 8589 of the Intl. Organization for Standardization (ISO 1988). Testing was performed at room temperature (20 ◦ C) with an appropriate and adequate artificial lighting, simulating daylight. Statistical analysis A two-way analysis of variance (ANOVA), followed by Tukey HSD test for multiple comparisons, and a Principal Component Analysis were carried out on the experimental data by the XLStat software (Addinsoft SARL, New York, N.Y., U.S.A.). For the ranked consumer test data, testers attributed to each descriptor a score equal to 1 for the “extremely liked” taralli, 2 for the taralli “liked very much,” 3 for the “moderately liked” taralli, 4 for the “slightly liked” taralli. Differences in the preference rank sums between all possible pairs of products were considered. Should any of these (absolute) differences exceed a critical value, the preferences for that pair of products differ from one another at the stated statistical significance level (n = 80, P ≤ 0.05, critical value = 40.2) (Basker 1988).

Results and Discussion The analytical characteristics of the flour used in the experimental trials are hereafter described: moisture 12.2%, ashes 0.51% d.m., protein (total nitrogen × 5.7) 11.7% d.m., dry gluten 10.6% d.m., gluten index 61. Table 1 shows the fatty acid composition and the oxidative and hydrolytic degradation of the lipid fraction of flour and of the

oils used in taralli making. The fatty acid composition of wheat flour lipids and vegetable oils corresponded to the one usually detected in these kinds of products (Uzzan 1996; Wuidart 1996; Prabhasankar and others 2000). In particular, the lipid fraction of flour was richer in polyunsaturated fatty acids (C18:2 ), whereas EVOO, OO, and OPO contained more monounsaturated fatty acids (C18:1 ) and RPO was characterized by short and medium chain fatty acids (C12:0 -C16:0 ). As for the polar compounds (PC), separated by silica gel column chromatography and composed by all the substances having higher polarity than the one of the unaltered triacylglycerols, they give a first indication of the level of degradation of a fatty matrix. Data reported in the table point out that the lipid fraction of flour show a degradation level definitely higher than RPO and above all than EVOO, OO, and OPO. The HPSEC analysis of the PC of wheat flour lipids showed values of diacylglycerols (DAG) and oxidized triacylglycerols (oxTAG) equal to 7.26% and 3.80%, respectively. The high content of ox-TAG, products of the primary oxidation, is probably due to both the conditions of the milling process and the high

content of polyunsaturated fatty acids in cereals. The oxidative and hydrolytic degradation of vegetable oils (EVOO, OO, OPO, and RPO) is comparable to those expected for these kinds of oils (Caponio and others 2005, 2009; Gomes and Caponio 1997, 2001). EVOO shows very low levels of triacylglycerol oligopolymers (TAGP), products of secondary oxidation; in fact, these substances are formed as result of the high temperatures of the refining process (Gomes and Caponio 1998; Gomes and others 2003). In Table 2, the mean scores and the results of the statistical analysis of consumer test are reported, as well as the definition of the descriptor used for the sensory evaluation. The analysis of data suggests that, with regard to the visual appearance, taralli made with EVOO were significantly more appreciated than those made with RPO and OPO, while for the odor the products made with EVOO and OO were much more appreciated than those prepared with RPO. Other descriptors registered no significant differences. Results emphasize that the replacement of refined oils, generally used in these kinds of formulations, with extra virgin olive oil—characterized by the presence of polyphenolic substances that

Table 1 –Fatty acid composition and oxidative and hydrolytic degradation of the lipid fraction of flour and oils used in taralli making. Parameters

Flour

Fatty acid composition (%) C12:0 nd C14:0 0.09 C14:1 nd C16:0 18.31 C16:1 0.13 C17:0 nd nd C17:1 C18:0 1.01 C18:1 11.49 C18:2 64.43 C18:3 4.43 C20:0 0.11 C20:1 nd Chemical determinations FFA (%) 7.49 PV (meq O2 /kg) 7.1 5.95 K232 K270 1.62 TAGP (%) nd ox-TAG (%) 3.80 DAG (%) 7.26 PC (%) 19.52

EVOO 1st batch 2nd batch

OO 1st batch

2nd batch

1st batch

OPO 2nd batch

1st batch

RPO 2nd batch

nd nd nd 9.03 0.32 0.03 0.04 2.06 78.80 8.17 0.88 0.67 nd

nd nd nd 10.73 0.52 0.04 0.06 1.75 79.58 6.17 0.89 0.26 nd

nd nd nd 14.30 1.50 0.06 0.11 2.40 69.07 11.41 0.82 0.31 0.01

nd nd nd 12.33 1.10 0.09 0.13 2.05 71.91 11.19 0.83 0.34 0.02

nd nd nd 12.03 0.79 0.08 0.11 1.96 73.04 10.92 0.76 0.26 0.04

nd nd nd 12.66 0.91 0.06 0.13 2.20 72.13 10.68 0.81 0.37 0.05

0.39 1.06 0.04 37.04 0.34 0.09 0.02 3.61 43.74 13.05 0.30 0.29 0.03

0.45 1.12 0.04 35.98 0.23 0.08 0.02 3.51 46.13 12.06 0.33 0.04 0.01

0.39 8.6 1.53 0.12 0.01 0.41 1.10 2.22

0.31 7.6 1.27 0.18 0.03 0.52 2.14 3.27

0.36 6.7 1.64 0.48 0.28 0.73 2.73 4.38

0.27 5.9 2.14 0.73 0.45 0.77 5.31 7.14

0.32 4.2 2.96 1.26 0.62 1.08 4.09 6.42

0.21 2.9 3.19 1.25 0.81 0.95 5.54 7.78

0.39 1.0 1.94 0.76 0.46 0.52 7.19 8.77

0.11 0.0 2.25 0.76 0.72 0.55 11.03 12.69

EVOO = extra virgin olive oil; OO = olive oil; OPO = olive-pomace oil; RPO = refined palm oil; FFA = free fatty acids; PV = peroxide value; K232 = specific absorption at 232 nm; K270 = specific absorption at 270 nm; TAGP = triacylglycerol oligopolymers; ox-TAG = oxidized triacylglycerols; DAG = diacylglycerols; PC = polar compounds; nd = not detected.

Table 2–Mean valuesA of consumer test and results of statistical analysis (n = 80) of taralli prepared using different oils. Descriptor and definition

EVOO

OO

OPO

RPO

Visual appearance Assessment of color, surface texture of sample Odor Intensity of odor smelt immediately above the sample’s breaking point Taste Taste intensity perceived after chewing the sample 4 to 5 times with closed mouth Friability Sample’s easiness to break into small fragments at the beginning of mastication

1.86

b

2.35

ab

2.39

a

2.46

a

1.87

b

2.08

b

2.32

ab

2.61

a

1.98

a

2.34

a

1.84

a

2.20

a

1.77

a

1.87

a

1.83

a

1.70

a

Overall acceptability Assessment resulting from the different attributes of smell, taste, and texture

2.16

a

2.18

a

2.04

a

2.49

a

A For the ranked consumer test data, the tester attributed to each descriptor a score equal to 1 for the “extremely liked” taralli, 2 for the taralli “liked very much,” 3 for the “moderately liked” taralli, 4 for the “slightly liked” taralli. EVOO = extra virgin olive oil; OO = olive oil; OPO = olive-pomace oil; RPO, refined palm oil. Different letters in row mean significant difference at P ≤ 0.05.

Vol. 76, Nr. 4, 2011 r Journal of Food Science C551

C: Food Chemistry



C: Food Chemistry

usually confer a characteristic sensory impression (Caponio and others 1999; Angerosa and others 2004)—does not affect the friability and taste of taralli but makes consumers appreciate much more their visual appearance and odor. Table 3 reports the results of the two-way ANOVA performed on the analytical data. The ANOVA models resulted to be significant for all the parameters with the exception of FFA. Moreover, it is possible to observe that the type of oil variable proved to have a stronger effect on the indices than the taralli making variable. In fact, the type of oil variable resulted to be significant for all the parameters with the exception of FFA, while the taralli making variable was significant only for PV, K270 , and ox-TAG. Tables 4 and 5 show the results of the two-way ANOVA, followed by the Tukey HSD test, of data regarding the oxidative and hydrolytic degradation of samples analyzed. It can be seen that the production process of taralli (Table 4) led to an increase in the oxidative degradation of the fatty fraction compared to the initial fat content (calculated by taking into account the contribution of different raw materials and the average fat content of the flour being equal to 1%), with a significant increase in PV, K270 , and ox-TAG. The hydrolytic degradation and the content of trans isomers were not significantly affected by the production process. The total data obtained confirm what Caponio and others (2009) already observed in a previous research on taralli prepared with refined palm oil. This indicates that the degradation of the fatty fraction, as a result of the technological process, is independent on the type of oil used. Regarding to type of oil variable (Table 5), the lipid fraction of samples made with EVOO presented levels of PV significantly higher than those found for RPO, while the determination of FFA did not lead to results being statistically different for the 4 types of oil used in the trials. However, it is known that these determinations can be successfully used to assess the level of oxidative and hydrolytic degradation of extra virgin olive oils. For other types of oil (OO, OPO, RPO) arising from a refining process, these indices do not offer any guarantee of assessing the real extent of oxidative and hydrolytic degradation, since FFA are removed during the neutralization phase and hydroperoxides are degraded during refining, due to the catalytic activity of the bleaching earth and the high temperatures reached during deodorization (Hui 1996; Capella and Fedeli 1997). The fact that no significant differences in terms of PV were found between OO, OPO (oils derived from a refining process) and EVOO can be explained by the fact that virgin olive oil, different from the lampante oil, is added to refined olive oil and refined olive-pomace oil to make them marketable (Official Journal of the European Communities 2001).

Taralli made with EVOO showed significantly lower values of K232 and K270 , which indicate significantly lower values of conjugated dienes and trienes, respectively, compared with taralli obtained using OO, RPO, and OPO. Furthermore, the use of EVOO led to significantly lower amounts of TAGP with respect to the other kinds of oil and to a lower amount of ox-TAG compared to OPO. The content of DAG was higher in RPO samples with decreasing values from OPO, OO, and to EVOO, although significant differences were found only comparing RPO with EVOO. Mistry and Min (1988) reported that DAG have pro-oxidant effects. Hence, as for TAGP, also DAG levels can affect the food quality. The evolution of PC was influenced by the one of the main classes of substances constituting them. TAGP can be effectively used to measure the secondary oxidative degradation of oil as they are stable and are not influenced by the conditions adopted; Table 4 –Results of two-way ANOVA, followed by Tukey HSD test for multiple comparison, of the obtained data in function of taralli making variable. Parameters

SF

FFA (%) PV (meq O2 /kg) K232 K270 TAGP (%) ox-TAG (%) DAG (%) PC (%) trans isomers (%)

0.64 4.7 2.30 0.74 0.38 0.84 5.00 7.20 0.30

T0 A B A B A B A A A

0.69 5.6 2.54 0.91 0.40 1.03 5.13 7.63 0.34

A A A A A A A A A

FFA = free fatty acids; PV = peroxide value; K232 = specific absorption at 232 nm; K270 = specific absorption at 270 nm; TAGP = triacylglycerol oligopolymers; ox-TAG = oxidized triacylglycerols; DAG = diacylglycerols; PC = polar compounds. SF = starting fat; T0 = extracted fat from taralli after production. Each value is the mean of 8 determinations and different letters in row mean significant difference at P ≤ 0.05.

Table 5 –Results of two-way ANOVA, followed by Tukey HSD test for multiple comparison, of the obtained data in function of type of oil variable. Parameters FFA (%) PV (meq O2 /kg) K232 K270 TAGP (%) ox-TAG (%) DAG (%) PC (%) trans isomers (%)

EVOO 0.69 7.5 1.62 0.28 0.05 0.76 1.96 3.79 0.00

a a c c c b b c b

OO 0.71 5.9 2.22 0.73 0.33 0.98 4.23 6.62 0.43

OPO a b b b bc ab b bc a

0.65 5.2 3.44 1.40 0.67 1.26 5.09 8.03 0.48

a b a a a a b ab a

RPO 0.69 1.9 2.40 0.88 0.53 0.72 8.97 11.22 0.47

a c b b ab b a a a

FFA = free fatty acids; PV = peroxide value; K232 = specific absorption at 232 nm; K270 = specific absorption at 270 nm; TAGP = triacylglycerol oligopolymers; ox-TAG = oxidized triacylglycerols; DAG = diacylglycerols; PC = polar compounds. EVOO = extra virgin olive oil; OO = olive oil; OPO = olive-pomace oil; RPO = refined palm oil. Each value is the mean of 4 determinations and different letters in row mean significant difference at P ≤ 0.05.

Table 3–Results of the two-way ANOVA performed on the analytical data. Model Determinations FFA (%) PV (meqO2 /kg) K232 K270 TAGP (%) ox-TAG (%) DAG (%) PC (%) trans isomers (%)

F

P-value

F

1.226 26.011 22.975 33.264 7.576 7.222 5.123 5.128 6.027

0.387