FERMENTATION OF GREEN OLIVES USING STARTER CULTURE AND QUANTITATIVE DESCRIPTIVE ANALYSIS FERMENTAÇÃO DE AZEITONAS VERDES UTILIZANDO CULTURA STARTER E ANÁLISE DESCRITIVA QUANTITATIVA

Resumo: Este trabalho tem como objetivo avaliar as características sensoriais, utilizando a análise descritiva quantitativa (QDA), de diferentes formulações de azeitonas verdes fermentadas por 120 dias com Lactobacillus plantarum ATCC 8014 e/ou culturas espontâneas (microbiota indígena). A fermentação foi realizada com solução salina em diferentes concentrações de sal (4 a 12%) e sacarose (0,1 a 0,7%). Foram avaliadas as análises químicas da solução de salmoura (pH, ácido lático e cloretos) e microbiológicas das azeitonas fermentadas. Os atributos organolépticos das azeitonas fermentadas foram realizados por painel sensorial treinado em relação aos atributos visual, olfativo, cinestésico e paladar. Os resultados mostraram que a equipe treinada conseguia se lembrar e descrever atributos consensuais. Este estudo revelou que a fermentação de azeitonas com cultura espontânea com sacarose a 0,7% e NaCl a 12% apresentou os escores mais altos em termos de aroma, odor, crocância e menores escores no sabor amargo. Palavras-chaves: Lactobacillus plantarum, Spontaneous, fermentação de azeitonas, análise descritiva quantitativa. Abstract: This work aim to evaluate the sensory characteristics, using the quantitative descriptive analysis (QDA), of different formulations of green olives fermented per 120 days with Lactobacillus plantarum ATCC 8014 or spontaneous (indigenous microbiota) cultures. The fermentation was undertaken with brine solution in different concentrations of salt (4 to 12%) and sucrose (0.1 to 0.7%). The chemistry analysis of the brine solution (pH, lactic acid and chlorides) and microbiological of the fermented olives were evaluated. The organoleptic attributes of fermented olives were assessed by trained sensory panel in relation to the attributes visual, olfactory, kinesthetic and taste. Results showed that the trained team could remember and describe consensual attributes. This study revealed that the olives fermentation with spontaneous culture using 0.7% sucrose and 12% NaCl, presented the high scores in terms of aroma, odor, crispness and lowest scores for the taste bitter.


Chemical and microbiological determinations
After 120 days of olives fermentation the following chemical characteristics were studied in the brine: pH was measured using a pHmeter (Hanna -HI 2212), titratable acidity is expressed as grams of lactic acid per 100 mL of brine and chlorides were analyzed by the Volhard method (GARCÍA; BALBUENA; FERNÁNDEZ, 1991). All the analyses were performed in triplicate.
The microbiological quality of olives at the end of the fermentation process was evaluated in relation to pathogenic Salmonella sp., and microorganisms indicators of microbial contamination (enumeration of thermotolerant coliforms, Staphylococcus aureus and sulphite reducing clostridia count), according to Brazilian legislation -Normative Instruction No. 30 (MAPA 2018).

Sensory analysis
The samples were profiled using QDA (STONE, 2012). Responsive to all the sensory properties of a product, the QDA methodology provides a complete word description for all product's sensory properties. To ensure that product is fully explained, it was necessary select and train the panel with session's of forty-hour.
Selection procedures, training, monitoring of assessors, definition of terminologies, ideal measuring range and evaluation of results were developed according to the standards used in descriptive terms and preparation of samples for training and memory, summarized a few propositions described by international standards (MAPA 2018;COI 2011;ISO 2012), previous research studies (APONTE et al., 2010;GALÁN-SOLDEVILLA;RUIZ PÉREZ-CACHO, 2012;GONZÁLEZ et al., 2007) and adapted to the availability of reagents and local products. Revista Tecnológica -Universidade Estadual de Maringá -ISSN 1517-8048 DOI: 10.4025/revtecnol.v29i1.51099 _______________________________________________________________________ The study protocol followed the ethical principles of sensory laboratory, approved by the ethical committee of the Regional University of Erechim, under the CAAE 03152512.7.0000.5351 registration number.
Panel of judges: The panelist's recruitment (67 judges) was carried out a personal contact with questionnaire and signature of consent. Candidates for judges were excluded if not have time for train and not like olives. The selection of the best candidates was carried out by discriminative tests (triangle test for difference testing): basic tastes (sweet, salty, bitter and acid), intensity of taste, mixture of basic tastes, odor and aroma recognition. The candidates that scored at least 80% of the correct answers were selected.
Sensory evaluation of olive fruits: For descriptive terms, the judges reviewed the sensory definitions, from pre-established standards, as similarities and differences between the olives offered for visual characteristics (appearance and color), olfactory (odor and characteristic olive aroma), kinesthetic (crispness and hardness) and taste (salty, bitter and acid). Thus, after defining the characteristics of the detailed description and setting up the sensory profile.
The references identified were presented to each judge. Specific train sessions were carried out for each attribute, which was assessed using a 9 cm unstructured scale (NBR14141, 1998). The extremes were defined in terms of minimum and maximum intensity of each specific attribute.
Four samples, with triplicates, were evaluated by 10 judges trained (age range 18-25 years). Among them, six women and four man that not smoke and not present history of allergies. Tests were performed in individual cabins with white light. Samples were served in plastic cups coded with three-digit numbers random for analyzes.

4 Statistical analysis
The panel leader introduced the evaluation data and verified the performance of judges with the following tests: three representative samples of each descriptor were presented to the panel, and each assessor evaluated three repetitions of three samples. The individual results of assessors and of descriptors (acid, bitter, appearance, aroma, color, crispness, hardness, odor and salty) were statistically analyzed by analysis of variance (ANOVA) and Tukey's test at 5% significance level. Analysis of variance of 2 factors (sample and repetition) was performed for each panelist with respect to each attribute and the judges (Table 1) were selected according to values of significant Fsample (p < 0.30) and non-significant Frepetition (p>0.05). After was evaluated the concordance of the judges with the group. The judges were then selected based on their ability to discriminate among different samples, repeatability and agreement with the group.
The analysis was performed used the software Statistica 5.0 (StatSoftInc®, USA) with 95% significance level. Correlations between variables were established by correlation analyses with the XLSTAT (2014.5.03) statistical software package.

Selection of olives formulations for sensory analysis
The brine solution of green olives fermented with spontaneous or L. plantarum cultures were evaluate in relation to pH, chloride and lactic acid ( Table 2). All formulations showed significant difference (p<0.05) in the pH values. As described by (CODEX ALIMENTARIUS, 1981;COI, 2011COI, , 2017, the final pH of naturally fermented olives must be equal or lower than 4.3. Only treatments S1, S3, S4, L2, L3 and L4 presented pH values less than 4.3, and thus complied with those standards. The lactic acid in the brine solution in the formulations S1, S4, L2, L3 and L4 maintained the levels equal or above the minimum recommended by Consejo Oleícola Internacional (COI, 2011) which is 0.3% of free acidity expressed as lactic acid (g/100mL). In addition, was observed a significant difference (p<0.05) of the higher lactic acid value of (L4) in relation to the other formulations. For the chlorides is possible to observe that the formulations with 12% of NaCl (S2, S4, L2, and L4) do not present significant difference (p<0.05) among them.
In relation of the microbiological analysis, the Salmonella sp. was investigated due to the potential of food hazard. None of the analyzed formulations presented Salmonella sp. The absence of coliforms is a good indicator of the appropriate sanitary conditions of the food. Thermotolerant coliforms were detected in the formulations S1 and S3, with higher counts of 10 2 UFC/g. These samples should be considered unacceptable and improper for consumption. The Brazilian legislation, not establish limits for sulphite reducing clostridia. As some treatments contained 4% NaCl, and this condition can allow undesired growth of these microorganisms, the analyses were performed in olives samples and the results no showed viable colonies. Moreover, the counts of S. aureus in all samples were lower than the limit (5 x 10 2 CFU/g) established by the RDC No.12 (ANVISA, 2001).
In this way, considering the criteria of: pH lower than 4.3, lactic acid content above of 0.3%, according to Anvisa (2001), and also the results of microbiological analyses, visual and Revista Tecnológica -Universidade Estadual de Maringá -ISSN 1517-8048 DOI: 10.4025/revtecnol.v29i1.51099 ___________________________________________________________________________ olfactory aspects, the formulations S1 and S3 did not attends the microbiological criteria, and S2, S3 and L1 did not meet the prerequisites of pH and lactic acid. Thus, these formulations were excluded of sensory analysis and the formulations chosen for further analysis were L2, L3, L4 and S4.

Sensory profile of the olives samples
For the sensory panel initially 67 judges participated in the recruitment and selection stages, and 16 judges developed a descriptive terminology (visual and descriptive analysis), olfactory (odor and characteristic olive aroma), kinesthetic (crispness and hardness) and taste (salty, bitter and acid) defining the similarities and differences among the sample, and 10 judges were considered trained for the QDA of fermented olives.
Variability between samples and attributes was observed comparing the sensory profiles of fermented olive (Table 3 and Fig. 2). The formulation L3 differs significantly (p < 0.05) considering the crispness and salty attributes. Moreover, the hardness in the formulation L3 did not differ from L2, but differed of the formulations L4 and S4. The formulation L3 showed higher average score for bitterness and lower for the crispness, hardness and acid, when compared with other formulations. These differences can be related to the low concentration of NaCl used in these formulations, because the salty taste can be suppressed bitter taste due to the use of lower concentrations of salt. In addition, in a previous study (ROSA et al., 2016a) verified that L. plantarum showed high counts, reaching fast adaptation at the beginning of fermentation (9 days), different of the spontaneous culture that require 20 days. Also is possible to observe that the trained team could remember and describe consensual attributes, according of the homogeneity of numerical values and low values of standard deviation.
Some authors provided a broader perspective to the use starter cultures in fermentation process; ensure safe and best sensory characteristics. However, the presence of spontaneous species result in special sensory character to the olives, which can be inhibited by the use of starter culture (APONTE et al., 2010;RANDAZZO et al., 2012RANDAZZO et al., , 2014ZAGO et al., 2013). These results corroborate with those of the present study, where the formulation S4 (12% w/v NaCl, 0.7% w/v sucrose and fermentation with spontaneous culture) was scored as the better formulation by sensorial analysis.
Mourad and Nour-Eddine (2006), observed in naturally fermented green olives, in the first fermentation stage (15 days), wherein was characterized by the growth of heterogeneous microorganisms, including aerobic bacteria, coliforms, staphylococci, lactic acid bacteria, lactobacilli, enterococci, psychrotrophs, lipolytic bacteria and yeasts, and after 60 days lactic acid bacteria and yeast were the predominate population in relation to the initial heterogeneous bacterial flora. Ruiz-Barba and Jiménez Díaz (2012) cited that spite of some available commercial inoculums, the use of starter cultures not have a habitual practice in the olive processing industry. However, it is possible to obtain more effective starter cultures with detailed selection of strains and relevant technological characteristics.

Conclusion
After the training steps, 10 tasters were considered trained to evaluate the QDA of fermented olives for attributes visual characteristics (appearance and color), olfactory (odor and characteristic olive aroma), kinesthetic (crispness and hardness) and taste (salty, bitter and acid).