You are on page 1of 10

International Journal of Food Microbiology 257 (2017) 31–40

Contents lists available at ScienceDirect

International Journal of Food Microbiology
journal homepage:

Impact of Saccharomyces cerevisiae and Torulaspora delbrueckii starter MARK
cultures on cocoa beans fermentation
Simonetta Visintina, Lacerda Ramosb, Nara Batistab, Paola Dolcia, Freitas Schwanb,⁎,
Luca Cocolina,⁎
University of Torino, DISAFA - Microbiology and Food Technology sector, Grugliasco, (TO), Italy
Universidade Federal de Lavras (UFLA), Departamento de Biologia, Lavras, (MG), Brazil


Keywords: Aim of this work was to study the impact of mixed cultures of Saccharomyces cerevisiae and Torulaspora del-
Cocoa beans fermentation brueckii and T. delbrueckii monoculture on the fermentation process conducted on two different cocoa hybrids,
Starter culture PS1319 and SJ02, in Bahia, Brazil. This was performed throughout studying physico-chemical changes during
Rep-PCR the fermentation process and analyzing volatile compounds and sensory analysis of chocolates. (GTG)5-PCR
fingerprinting was used to type isolates at strain level allowing to assess the implantation of the starter cultures
Volatile compounds
added. Resulted clusters were composed by T. delbrueckii strains isolated during the first 24 h of fermentation.
On the contrary, S. cerevisiae, the most strongly fermenting ethanol-tolerant species, took over the fermentation
at a second stage. Quantification data of T. delbrueckii during spontaneous fermentation confirm the attitude of
this species of not being so commonly involved in this process.
This study also showed that the inoculum influenced the PS1319 hybrid end-product quality, changing
analytic profile and sensory perception of chocolates. No big influences were recorded for SJ02 hybrid, but this
may be improved. In combination with S. cerevisiae, T. delbrueckii had a positive influence on the analytical
profile of chocolates. The application of starter cultures did change the aroma profile of the resulting chocolate
as determined by GC–MS; in some case the differences observed had a significantly impact on the consumer
perception of the chocolates.

1. Introduction chocolate (Ramos et al., 2014). Fermentation process involves complex
microbial activities and biochemical changes that have been recently
In Brazil, like in some other part of the world, within the farmers it deeply reviewed by Schwan and Fleet (2014). The microorganisms re-
is common to cultivate different cocoa varieties in the same field in sponsible for the fermentation are yeasts, lactic acid bacteria (LAB) and
order to prevent destruction of the cocoa trees by fungus. Therefore, acetic acid bacteria (AAB), moreover species of Bacillus sp., other bac-
beans from different variety are collected and spontaneously fermented teria and filamentous fungi could also grow with consequent influence
in the same boxes. This procedure can affect the quality of the chocolate on quality of the process. Recently, Ho et al. (2014) elucidated the
as recently reported by Ramos et al. (2014), indeed the variability due importance of yeast growth and activity for successful cocoa bean fer-
to cocoa variety is an important factor influencing end-products quality, mentation and their involvement on the development of chocolate
together with the fermentation process itself (Schwan and Wheals, aroma.
2004). Torulaspora delbrueckii is not normally involved in cocoa beans
PS1319 and SJ02 are cocoa varieties produced by CEPLAC fermentation, indeed this species is not regularly found worldwide. Its
(Comissão Executiva do Plano da Lavoura Cacaueira, Bahia, Brazil). presence has been recorded in few cases in Ghana, Ivory Coast,
These varieties show high productivity and Moniliophthora perniciosa Malaysia and Ecuador (Jespersen et al., 2005; Nielsen et al., 2007;
resistance (Mandarino and Sena Gomes, 2009). Papalexandratou et al., 2013; Papalexandratou et al., 2011; Visintin
The cocoa fermentation is still traditional and uncontrolled and an et al., 2016). However, this is the first time this yeast is used as starter
adequate process together with an appropriate selection and handling culture for cocoa beans fermentation as monoculture and in mixture
of raw materials, drying and roasting can influence characteristics of with Saccharomyces cerevisiae.

Corresponding authors.
E-mail addresses: (F. Schwan), (L. Cocolin).
Received 14 January 2017; Received in revised form 1 June 2017; Accepted 6 June 2017
Available online 07 June 2017
0168-1605/ © 2017 Elsevier B.V. All rights reserved.

Additionally. and Tods L2 (5′-CAAAGTCATCCAAG- each were collected at 0. cerevisiae strain showed the best resistance profile for all μL. 144 and 168 h of the CCAGC-3′) and Tods R2 (5′-TTCTCAAACAATCATGTTTGGTAG-3′) for fermentation process. Milli-Q water. delbrueckii starter from the surface of the center of the fermenting cocoa mass. S. T. Direct extraction of DNA from cocoa samples main crop of 2014 (September–December). The purified isolates were stored at − 20 °C in YPD broth containing 25% of glycerol. point on the calibration curve was measured in triplicate. The final volume of 25 mL of diluted pulp broth [10 g/L Yeast extract (Merck. physico-chemical time. colonies of each type was re-streaked and purified. Metabolites extraction and HPLC analyses Twenty-five grams of cocoa beans and adhering pulp were added to The carbohydrates. 2000. particularly β-glucosidase. For standard curves. Visintin et al. cerevisiae and T. using the Dice similarity coeffi- cient and dendrograms were obtained by means of UPGMA clustering The fermentation experiments were conducted at the Vale do algorithm. delbrueckii. This was incubated at 30 °C for 3 (yeasts) to 5 days (AAB).. delbrueckii monoculture on fermen. and the beans were immediately For DNA extraction. Serial 10 g of cocoa beans of each sample as described by Rodriguez-Campos dilutions in peptone water were prepared and yeasts were enumerated et al. 2014). Moreover. tential starter culture strains. All analyses were performed in tripli- that is not systematically found in cocoa beans fermentation. 2000). Whitehouse Station. Himedia).. this 2. Specific primers used in this study were previously described by hybrids were also spontaneously fermented (without inoculation). cerevisiae ID67 in co-culture natant was separated from the precipitate. S. alcohols and organic acids were extracted from 100 mL of peptone water (Himedia) and homogenized manually. 2016). and cen- delbrueckii ID103 monocultures. 20 g/L agar. 4 °C). delbrueckii. Milan. 2. S. analyzing the volatile com. et al. 30 g/L calcium carbonate. cerevisiae and T. International Journal of Food Microbiology 257 (2017) 31–40 Aim of this work was to study the impact of mixed cultures of S.. 10 g/L yeast extract. 1994) as previously described by Dal Bello et al. qPCR 106 cells/g and 105 cells/g of cocoa for S. Plates were tation process conducted on two different cocoa hybrids. Juliana cocoa farm in Igrapiúna. The samples were taken approximately 40 cm T.06 m3 wooden boxes. The pH values and counted using a Neubauer chamber and DNA was extracted as pre- temperatures were evaluated during fermentations by an average of viously described and serially diluted (1:10) from 108 and 107 for S. glicosidase. vortexed for 5 min and the supernatant collected in a proximately 3 h after the breaking of the pods and was performed in Falcon tube. and the super- with inoculation of hybrid PS1319 with S. The cocoa pods were manually opened with a machete. DNA extraction from pure cultures 2. ZH. Switzerland) following the method de- moreover the latter was inoculated in lower amounts due to the fact scribed by Batista et al.. cerevisiae and T. of sterile peptone water (1 g/L Peptone. The cells were re.2. The by spreading on YPD agar containing 100 μg/mL chloramphenicol final volume of 25 mL of diluted pulp was centrifuged. The precipitate was re-sus- with T. Thermo Fisher Scientific. agar (50 g/L glucose. the parameters used for screening such as temperatures. based on previous investigations (Visintin Thermo Fisher Scientific.7. Brazil). Materials and methods 2. India). This solution was spread over the cocoa beans reaching a concentration of approximately 2. pH. The cells were sterile plastic pots and transferred to the laboratory. Italy). (2014). standard deviation were calculated). Starter cultures DNA was extracted from 1 mL of an overnight culture of yeasts as S. The step was repeated and the tubes containing 20 mL of 0. Total DNA Peptone (Himedia. cerevisiae. All Zott et al. Merck). isolated and subjected to repetitive extragenic tolerance screening with the only exception of ethanol concentration. The square root of the number of pounds and by sensory analysis of chocolates. vortexed. protease and At least 5 yeast colonies for each sample at each sampling point pectinase. (2011) and successively modified by Ramos et al. 10 min. Real-time PCR was carried out using the Rotor-Gene Q System brueckii to accelerate the fermentation process (Ramos et al. delbrueckii IC103 also exhibit a suitable profile for the were randomly selected. 10 g of each sample were diluted with 10 mL of transferred to the fermentation house. three different points into the fermentation boxes using a portable pH cerevisiae and T. by using the Nanodrop Instrument (Spectrophotometer ND-1000. 72. delbrueckii ID103 and for hybrids PS1319 and SJ02 with T. Microbiological analysis 2. Each meter Q400HM (Quimis. Italy) and then standardized at 100 ng/ et al. (2010). (2015). (Qiagen.. DNA was quantified by using the Nanodrop Instrument covered by centrifugation (7000 rpm. and 2 mL of (Merck) in triplicate and AAB were determined by spreading on GYC supernatant was filtered through a 0. Mumbai. The inoculum was prepared in YPD trifuged as described above. 20 g/L was centrifuged and the supernatant divided from the pellet. respectively.4.8. down to 102 cell/mL. Bahia. (2010) and Díaz et al. Milan. 2.1. 20 g/L dextrose (Merck)] at 30 °C was extracted from 100 mg of pellet as previously described (Cocolin at 150 rpm and replicated every 24 h for five days. pended in an additional 5 mL of Milli-Q water. SP. cerevisiae was used in higher population than T. Fermentations of the hybrids were performed solution were centrifuged (7000 rpm. 2016). (GTG)5-PCR fingerprinting cocoa-specific strain showed positive results for almost all the enzy- matic activities. Fermentation experiments and sampling strains through Bionumerics software. DNA was quantified spontaneous cocoa beans fermentation and they were selected as po.S. 48. delbrueckii IC103 strains were isolated from previously described (Cocolin et al. Brazil. placed in culture were cultivated in YPD agar at 30 °C for 24 h. The ripe cocoa pods from two different hybrids PS1319 and SJ02 were harvested during the 2. NJ). 10 min) and re-suspended in 1 L (Spectrophotometer ND-1000. The DNA profiles were compared with those of the inoculated 2. palindromic PCR (rep-PCR) to confirm the presence of starter culture. del. delbrueckii and T. cerevisiae IC67 and T. 24. 120. delbrueckii respectively. The two cate. 96. (Versalovic et al. Hombrechtikon.3.4%. it was one of the few strains that resulted positive for Rep-PCR was performed using the single oligonucleotide primer (GTG)5 esterase activity (Visintin et al.6. (2013) and are as follow: SC-5fw (5′- fermentations were evaluated for seven days.. After the incubation performed throughout studying microbiological.22 μm membrane (Millipore) for 32 . the colonies grown in each medium were counted (mean and changes during the fermentation process. Samples of approximately 100 g ATTCCCCT-3′) for S. The fermentation started ap. and the amount used for AGGAGTGCGGTTCTTTGTAAAG-3′) and SC-3bw (5′-TGAAATGCGAG- each assay was 300 kg of fresh beans. 2004). as well as ethanol concentrations and tolerance to osmotic stress.5. Merck) containing nystatin (0.

1 ± 0.4 ± 0.9 mm × 30 cm) was set up at 50 °C. Brazil). model LC-10Ai.9 ± 0b 4 ± 0.9.6 ± 0.2 4.1 na 3.2 SJ02 25.4 ± 0.5 ± 0b 6.1 2.1b 2. Germany).5 ± 0 5. 41% male and 59% female.9 ± 0.2a 40.1b 43.2 3.2 ± 0.1 6.1c 4.1b 2.1b na PS1319 T 24.4 na 2.1a 3.8 ± 0.2 4.4 ± 0.2 ± 0.3 ± 0. log10 CFU/g SJ02 T 5.1 ± 0.5 ± 0.2 ± 0 5.2 3.3 ± 0.8 ± 0.2 ± 0.8 ± 0.25 μm) (J & W Scientific.1b 30.1 5.1c 43. Volatile compounds were identified by comparing the retention linear curve in order to estimate the concentration of the compounds.3a 47. chromatography system (Shimadzu.1 3.3 ± 0.1c 48 ± 0.3 48.2 na PS1319 T 3.2 ± 0.3 ± 0 ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ Sig. (2015).4 6.6 ± 0b 25.2 4.5a 5.7 ± 0. Reten.4 ± 0.1 3.4 ± 0.1 6. T° and CFU are the mean ± standard deviation of three analyses. Different letters a-c within the same fermentation time and the same parameters are significantly different (p < 0.1 45.1 35.d.6 mL/min. The sample concentrations were calculated gradient of 60 °C to 230 °C at 10 °C/min.4 ± 0.7 ± 0c 5.6 ± 0.3 ± 0 5.4 ± 0.3 SJ02 4.2b 35.1 2.1 ± 0.1 2. EUA).5 ± 0c 7 ± 0c 7 ± 0 na PS1319 T 6.7 ± 0.2 48.1 ± 0 6.2a 25.9 ± 0.8 ± 0.1 ± 0.4 ± 0. The carrier gas (N2) was used at a flow obtained by injecting different concentrations of the standards at the rate of 1.3 ± 0b 5.9 ± 0 5. NS Yeasts. 2.3 ± 0.2 4.9 ± 0.6 ± 0.2 SJ02 3. na means not applicable.1b 2.2 ± 0. times of the compounds in the samples with the retention times of Standard (purity N 99.4 ± 0.1 na 2.2b 2.8 ± 0.2a 2. cerevisiae ID67 and T.3a 47 ± 0.4 ± 0. Each area was successively plotted in a 2 min.6 ± 0 3 ± 0 3 ± 0.6 ± 0.6 ± 0.8 ± 0 6. lowed by a check-all-that-apply (CATA) question.05).2 ± 0a 3.6 ± 0. tion time of the standards injected using the same conditions were used The temperature program began with 5 min at 60 °C. A Shimadzu ion exclusion column (Shim-pack compounds were analyzed using a Shimadzu GCMS QP2010SE model SCR-101H. NS NS NS NS NS Yeasts.7 ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎ Sig. Injections were performed by fiber exposition for same conditions of the samples. Shimadzu Corp.9.2 4. fructose.4 ± 0. SP. delbrueckii ID103.01 and p < 0.1b 4 ± 0. × 0.7 ± 0.0 g) were sumers evaluated how much they liked each sample using a 9-point 33 .5 ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎ ⁎⁎ Sig.1a 5.5 ± 0b 5.9 ± 0.2 3.7 ± 0.3 4.5 47. For the acceptance test.9a 36.2a 4.1 2.5 ± 0 3.8%).8 ± 0 4 ± 0.5 ± 0. USA).9 ± 0.7 ± 0.3 ⁎⁎⁎ Sig.25 mm i.2 24.2 ± 0.1 PS1319 6 ± 0b 6.3 ± 0 PS1319 5 ± 0a 4.5 ± 0.1. and lactic areas of all identified compounds.7 ± 0. and citric acid were standard compounds injected under the same conditions.6 ± 0 6. while the calibration curves were maintained at 230 °C for 15 min.1 ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ Sig.2a 5.1 ± 0.5 ± 0.2 ± 0.3 ± 0a 5.7a 46 ± 0.5 ± 0 3. ⁎⁎⁎ and NS that indicate significance at p < 0.2. glucose.8 ± 0. Germany).2 ± 0c 2.6 ± 0b 5.5 ± 0a 5.6 ± 0.5 ± 0. 7.1 46.1a 3.1a 4.6 ± 0.4 4.95 mL/min.2b 25.7 ± 0.9. temperature measurements.2 ⁎⁎ ⁎⁎ ⁎⁎ ⁎ ⁎ Sig.1b 4.4b 47.6 ± 0b 4 ± 0c 5.8 ± 0 24. The relative percentages of in- acid from Fluka Analyticals (Seelze.2 ± 0.2 ± 0 3. headspace (SPME-HS). Fermentation code Fermentation time (h) 0 24 48 72 96 120 144 168 192 A pH PS1319 ST 3. After the pulp extraction the cocoa beans were de. Volatile compounds extraction and GC–MS regular consumers of dark chocolate.8 ± 0.1c 41 ± 0.1 4.1 na 6. dividual compounds were calculated from the total contents of volatiles on the chromatograms.1a 46. The tests were con- ducted on 71 adults over 18 years of age.1a 4 ± 0.1c na PS1319 T 6.3 ± 0 3 ± 0 2. A 50/30 μm divinylbenzene/carboxene/polydimethylsiloxane Japan) provided with a dual detection system: a UV–Vis detector (DVB/CAR/PDMS) fiber provided by Supelco was used to extract vo- (SPD10Ai) for the detections of acids group (210 nm) and a refractive latile compounds.1 4. log10 CFU/g PS1319 ST 4.2 4 ± 0.2 4.4 ± 0.7 ± 0 ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ Sig.8 46. followed by a to identify the compounds.1 ± 0.2 4.4 ± 0. 2.3 ± 0a 6. Quantitative purchased from Sigma-Aldrich (Saint Luis.5 ± 0. respectively.2a 24.6 ± 0c 2. Sensory analysis The dried beans were sent for chocolate production at Sartori and Sensory analyses of samples made from controls and inoculated Pedroso Alimentos Ltda.1 4. ST means inoculated with S.1b 5. The by Batista et al. extracted using the solid phase micro extraction technique in the prived of the testa and crushed using a pestle and mortar.2 ± 0 3.2 46 ± 0.05.2 ± 0.7 ± 0.4 ± 0.3a 47.2 ± 0.1c 30. NS NS NS AAB. the HPLC analysis.6 ± 0. p < 0.1b 4 ± 0a 3 ± 0 2.3b 47.7 ± 0.1 2. CA. NS AAB.4 ± 0.7 ± 0.5 ± 0.2 40.8 ± 0.3 ± 0.2 ± 0b na 5. (Sao Roque. organic acids (acetic.S.2 4.6 ± 0 SJ02 5.9 ± 0. log10 CFU/g PS1319 ST 7 ± 0c 6. B pH SJ02 T 3. Folsom.7 ± 0.2 PS1319 3. Perchloric acid equipped with a capillary column of silica Carbowax 20 M (100 mM) was used as the eluent at a flow rate of 0.3 ± 0.2 ± 0.4 ± 0.2 ⁎⁎⁎ ⁎⁎⁎ ⁎ ⁎ Sig.5 ± 0.2b 4. yeast and AAB counts during spontaneous and inoculated PS1319 (A) and SJ02 (B) hybrids fermentations.1 6.8 ± 0.4 ± 0.2 4 ± 0.2 4.1a 5 ± 0.3a 4. Chocolate preparation and analyses 2.1 4 ± 0.7 ± 0.2a 5.5a 2.2 PS1319 25. NS NS NS NS T°C PS1319 ST 24.7 ± 0.2b 48.3 4. The carbo.1 4.1 ± 0.6 ± 0. (30 m × 0.2 3.3 47.9 ± 0.1b 4.2 34.001 and no significance.1a 4.1a 4..3 ± 0.9 ± 0 24. Samples were transformed into a fine powder hydrates (glucose and fructose).6 ± 0. Visintin et al. were bought from Merck (Darmstadt.8 ± 0.6 ± 0b 5. delbrueckii ID103.2 ± 0.1 ± 0. International Journal of Food Microbiology 257 (2017) 31–40 Table 1 pH. lactic and citric with a grinder (IKA A11 Basic analytical mill) using liquid nitrogen and acids) and alcohol (ethanol) analyses were then extracted using a liquid transferred to a sealed vial. ⁎⁎.7 36.3 ± 0.5 ± 0 4.1 2.1 3.2a 2.8 ± 0.2 ± 0. acetic acid and data of the identified compounds were obtained by integrating the peak ethanol.4 ± 0.1 2.1c 46.3 4. the temperature was then using an external calibration method. The values of pH. NS NS NS NS NS NS NS NS T°C SJ02 T 24.4 ± 0.7 ± 0.9 ± 0.7 ± 0.1 25 ± 0.7 ± 0b 5. NS NS Statistical data are expressed as ⁎. log10 CFU/g SJ02 T 6.4 ± 0. All samples were examined in duplicate.1 ± 0 6.6 ± 0.1 ± 0.7 ± 0. The fiber was balanced for 15 min at 60 °C and then index detector (RID-10Ai) for alcohols and carbohydrates as described exposed to the cocoa powder for 30 min at the same temperature.3 ± 0.1a 25. and T with T. The chocolate was fermentations were performed using a consumer acceptance test fol- prepared with 70% cocoa. the con- The volatile compounds of the chocolate samples (2.

1 4. the temperature of to 2. 72. delbrueckii ranged from (25.001 and p < 0. cerevisiae and T. deed.1 6. delbrueckii 2 ± 0. in the case of PS1319 inoculated with the mixed culture.2 3. 3.001 selected) at strain level allowing to confirm the implantation of the and p < 0. citrus flavor. delbrueckii 4.4 ± 0.3 4.1 4. ranging from 3 to 6.2 6. at day 0.7 ± 0. cerevisiae and T. pH values (p < 0. ranging from 4. T. measures (p < 0. In the control. The concentration values expressed as log CFU/g are mean ± standard deviation of three replicates.7 ± 0. cerevisiae and T.3 ± 0.4 ± 0.3 2. significant differences were re- propriate to describe the chocolate. Table 1B) and during chestnut flavor.9 ± 0.7 ± 0.7 ± 0.1 ± 0.05) were subjected to a post-hoc Duncan comparison to identify significantly different samples. delbrueckii ID103. Significant differences were also observed within PS1319 inoculated 3.3. banana.8 ± 0.2 4. rancid.10. delbrueckii population between PS1319 inoculated fermentations and the control one.2 and 46. to check Dynamic changes of the population of the two starter culture species for significant differences in the profiles. desirable bitter.998 and 0. groups higher in PS1319 experiment inoculated with S.999 and efficiency reaction values of 97 and 94% were obtained 3. In all three cases a strong de.7 ± 0. 1993). delbrueckii and the control one at 96.4 ± 0. The attributes were acid flavor.6 log cfu/g during the last phase. Table 1A) and (GTG)5-PCR fingerprinting was used to type 157 isolates (randomly during all the fermentations in the case of the AAB counts (p < 0.1 4. Visintin et al.3 2.2 3. 120.1 2. roast. The cluster analysis of the profiles obtained by (values ranging from 4. and T with T. Analyses were performed with during cocoa fermentations were monitored by qPCR.8 ± 0. fruity.8 ± 0.1 4. 120 and 192 h for detected (2 to 2.01). All fermentations lasted 192 h. were distinguished. temperatures and counts. Statistical analyses parameters e. Quantification of T. whereas in the control assay it was almost not spectively).4 2.5 ± 0. earthy flavor. showing the same bands profile as S.2 3.7 ± 0. cerevisiae and T. (2015). In all the cluster analyses.2 ± 0. Differences were recorded at 24.7 log cfu/g.7 log cells/g higher (30. In- to PS1319 inoculated with T. Standard curves IBM SPSS 21. cerevisiae 3.1 ± 0.4 3.3 log cfu/g).001) inoculated with S. Fermentation code Species Microbial counts expressed as log cell/g during fermentation time (h) 0 24 48 72 96 120 144 168 192 PS1319 ST S. 30. except at 120 h. as well as GC–MS could not be controlled. 120 and 168 h in the yeast counts (p < 0.5 ± 0.4 log cfu/g.1 5 ± 0. T. coffee flavor.3 ± 0. 72. buttery. as reported in Table 1A.1 4. undesirable bitter. cerevisiae and T. peanut almost all the process for AAB counts (p < 0. delbrueckii primers.1 ± 0.1 2. Yeast counts were significantly lower in PS1319 control starter cultures strains.2 na PS1319 T T. respectively.7 ± 0. 96 and 168 h in the yeast counts (p < 0.9 and 25. delbrueckii strains.1 6 ± 0. at all ranged from 4. S. cerevisiae 5.1 5. ex- flavor.2 4. delbrueckii profile. vanilla flavor and woody. the consumers were asked to observed after 24 h for pH values and at 0 and 72 h for temperature evaluate seven sensory attributes and select those they considered ap. 94% of the isolates grouped as T.6 ± 0 3 ± 0. the population of S. while S.4 ± 0. 96.2 2 ± 0. corded between inoculated and spontaneous fermentations at 0. a Below quantification limit.S. delbrueckii in PS1319 inoculated with S. however all assay were performed at same time data were analyzed using a one-way analysis of variance (ANOVA).1 5.4 to 6. delbrueckii and control fermentations at 0.7 ± 0 na PS1319 ST T.7 ± 0 5. than in the in- Table 1A and B. Furthermore. The lowest detection limit was 102 cells mL− 1. delbrueckii was significantly PS1319 (inoculated only with T. AAB counts were significantly in Fig.2 ± 0.1 ± 0 5 ± 0. 96. delbrueckii. The assays were performed in farm scale and these parameters Data obtained from pH. cerevisiae 34 . cerevisiae and T.1 ± 0. delbrueckii 4. cerevisiae ID67 and T.9 ± 0.1 ± 0. cerevisiae and T. Implantation of starter cultures with S. Results showing significance variance (p < 0. B and C.7 to 5. For the CATA question.1 4.1 7.2 5. T. The temperature. pH and counts of yeasts and AAB during PS1319 Concerning PS1319 experiments. delbrueckii were 3.1 4 ± 0 PS1319 S.3 ± 0.1 4 ± 0. delbrueckii ID103. del.3 ± 0.2 °C re.5 a a a a a PS1319 T. delbrueckii experiments are shown crease was observed later in the process.3 na. Results for S. delbrueckii) resulted in 4. respectively) with respect to and 6 log cells/g after 24.2 ± 0.1 °C.0001. delbrueckii by qPCR analysis While data obtained from sensory analysis were analyzed using a Friedman test and a post-hoc paired Wilcoxon rank sum test.7 to 5 log cfu/g in the first 48 h and then decreasing to 3 sampling times except at 192 h. 3. 5. 48. delbrueckii populations during inoculated PS1319 and SJ02 hybrids fermentations by qPCR. respectively. which lasted 168 h. and subject to the same conditions.2 ± 0.g.1 4.2 ± 0 5. At 48 and 96 h. significant differences were and Sidel. environmental temperature and microbial contamina- tion. delbrueckii 2. delbrueckii 4.3 and 41 ± 0. cerevisiae and T. not applicable.001).7 ± 0. Microbiological analysis of spontaneous and inoculated fermentations detected and quantified during inoculated and spontaneous fermenta- tions as shown in Table 2. 24.1 6.2 2.1. except oculated fermentation (inoculated with S. respectively. International Journal of Food Microbiology 257 (2017) 31–40 Table 2 Results of quantification of S. hedonic scale with 1 = dislike extremely and 9 = like extremely (Stone Concerning SJ02 hybrid fermentations. microbiota present in the different assay affect directly the parameters such as pH and temperature. variations may also be related to external 2. delbrueckii and to the control During SJ02 cocoa bean fermentations.2 5. 48 h and 72 h. correlation coefficient of 0.2 3 ± 0.8 ± 0. respectively.2. PS1319 and SJ02 inoculated with T.1 software package. 7 ± 0 and 7 ± 0 Log10 cfu/g) with respect added to the respectively inoculated batches.01 and p < 0.6 ± 0. sweet.2 °C. 144 and 168 h. PS1319 inoculated only with T. PS1319 inoculated with the mixture of S.6 ± 0. cerevisiae and T. delbrueckii.1 4. Fermentation code: ST means inoculated with S.5 ± 0. cerevisiae and T. 1A.2 to 5 ± 0 Log10 cfu/g) with respect to Rep-PCR for PS1319 inoculated with S. Quantification of S.1 a a a a a SJ02 T. The tests were cept for 96 and 192 h.5 ± 0.5 ± 0 4. and the inoculated boxes for the first 96 h. delbrueckii).6 ± 0. In PS1319 The temperature values were significantly different (p < 0. cerevisiae was and SJ02 spontaneous and inoculated fermentations are reported in lower in the control.5 5. 72. were established for each primer set. cerevisiae and T. brueckii (4 ± 0.001. cerevisiae and T. delbrueckii.4 ± 0. astringent. Although starter culture and consequently the performed as previously described by Batista et al.2 ± 0.7 ± 0.5 log cell/g.4 ± 0 6.01). delbrueckii was almost not detected as reported in Table 2.05).2 SJ02 T T.2 4.1 3.3 ± 0.

cerevisiae was present with 75 and 50% of the (Fig. delbrueckii (C) yeast isolates. cerevisiae and T. were calculated using Dice correlation index. increasing up to 100% at 48 and 96 h. In case of SJ02 monoculture experiment. in case of PS1319 inoculated 35 . delbrueckii inoculated in PS1319 in Isolates of those main clusters showed a similarity of 95 to 100% as monoculture corresponded to 91. isolates. obtained by means of cluster analysis of the Rep-PCR profiles.) represented 1% of the isolates. delbrueckii (B) and SJ02 inoculated with T. 1B). S. red and green lines indicate S. PS1319 inoculated with T. re. 1. cerevisiae ID67 and T. is reported. delbrueckii was present at day 0 with a total of 100% of the profiles while at day 3 and 4. delbrueckii profiles were spectively (Fig. composed only of strains isolated at day 0.S. 1A). B and C. with the respective day of isolation. delbrueckii ID103 inoculated strains. International Journal of Food Microbiology 257 (2017) 31–40 Fig. (For interpretation of the references to color in this figure legend. delbrueckii (A). 25 and 20% at days 0. cerevisiae and T. 1C). S1 and U1 in the strain code refer to S. Visintin et al. Clusters of T. the reader is referred to the web version of this article. 1A. Strain code. reported in Fig. delbrueckii isolates groups respectively. T. Dendrogram obtained by cluster analysis of PS1319 inoculated with S. T. respectively (Fig. The groups. 1 and 2.

Within this group 2-acetylpyrrole and acetoin resulted as the most impacting. as reported in Table 3. 2. G and I. isovaleric and 3-hydroxy-2. D.S. del- brueckii only (Fig. while during PS1319 fermentation in- 3. with S. 1. esters and lactones. Statistically different amounts of solerone was detected in chocolates made with PS1319 inoculated Fig.0 g/kg.01 g/kg of ethanol at 72 h.6. F. compared to the control (Fig.001) 2-phenylethanol content in PS1319 inoculated fermentations with respect to the control one. alco- hols. aldehydes and ketones. reaching the peak of ethanol production (around 7 g/kg) at 72 h and 96 h for boxes inoculated with S. lactic and citric acids). 2F). 2A. delbrueckii than in the control for PS1319 trials (Fig. Alcohols was the main group of volatile compounds detected. respectively. D and L) and at the end of the process (Fig. D and F). H and L show the concentration of carbohydrates. significant differences among these classes (except for pyrrole compound) were noticed. sugars (glucose and fructose) and alcohols (ethanol) of the cocoa beans sam- ples analyzed is reported in Fig.. Volatile compounds SPME-GC/MS method enabled the identification of 36 compounds in PS1319 chocolate samples and 34 in SJ02 trials.8 ± 0. C. E. Content of 2-phenylacetic acid was higher in chocolates made from T. delbrueckii inoculated cocoa beans clones. Carbohydrates were consumed faster in the fermentation in- oculated with S. On the other hand. no significant dif- ferences within SJ02 inoculated with T.methyl-4- pyrone acids were found to be significantly different (p < 0.8 e 1. delbrueckii (16. delbrueckii (0. as reported in Table 3. Chemical changes during fermentations Composition in organic acids (acetic. Different trends were observed for lactic acid within the samples as reported in Fig. cerevisiae and T. the maximum concentrations were higher between SJ02 inoculated fermentations and the control: 1. E. S. A peak of lactic acid was observed between 144 and 168 h (Fig. 1 and 2. ethanol and organic acids in the beans. On the contrary. delbrueckii and SJ02 inoculated with T.7%) with respect to the others. These compounds were mainly grouped into acid. hexanoic. Visintin et al.7 ± 0%) while it was not found in the control sample. Ethanol concentration in the pulp was higher in the PS1319 fermentation in- oculated with yeasts. C. delbrueckii (1. (Fig. 2E). 1A and C). G and I show the results obtained from the analysis of the pulp. SJ02 hy- brids fermentations showed the same value of 6. Within PS1319 clone experiments. except for SJ02 inoculated box. Fig. GC–MS data showed significantly higher amounts of aldehydes and ketones compounds in chocolate made from PS1319 inoculated with T. This is in agreement with the poor competitive abilities of this species in mixed yeasts ecosystems (Bely et al. delbreueckii. 2A and E). 3. 3. (continued) with S. 2B. cerevisiae and T. The concentration of carbohy- drates was similar in all the unfermented cocoa beans as shown in Fig. respectively. 2B. strains of T. Fig. Inoculated experiments showed the faster consumption of citric acid in the first hours as shown in Fig.4.2%) and PS1319 in- oculated with T. Sensory evaluation of chocolate oculated only with T. 2. 2008). Within the acids. The highest concentration of acetic acid was re- corded within 144 and 168 h. 2. cerevisiae and T. Citric acid was the higher detected compound present in all the experiments. delbrueckii were isolated at day 0. cerevisiae and T. delbrueckii.05) among PS1319 experiments. cerevisiae profiles were observed Sensory analyses allowed to differentiate between the chocolates 36 .1 ± 0. T. For the acetic acid. 2A. For what concerns alcohols group. del- brueckii box. respectively. cerevisiae and T.5. as shown in Table 3. 2B. delbrueckii and the control on the volatile compounds were observed. International Journal of Food Microbiology 257 (2017) 31–40 from 0 to 5 days in PS1319 inoculated with S. delbrueckii was only isolated at the beginning of the process and never found after 2 days of fermentation. statistical analysis showed a significantly higher (p < 0.

Discussion mentation (scored 6.S. during the fermentation process. 3A cerevisiae has been repeatedly defined as one of the most important yeasts involved in cocoa beans fermentations. Moreover. While S. Brazil. cerevisiae ID67 and T. delbreuckii showed a re- within the samples (scored 7. judged by the sensory panel. Changes of carbohydrates (glucose and fructose).001). sweet.1). fruity and roast. 2G and 2H correspond to SJ02 conducted with T. A and B correspond to PS1319 fermentation inoculated with S. Ramos et al. delbrueckii. The evaluation of the seventeen parameters used by the sensory The aim of this work was to study the impact of mixed cultures of S. SJ02 inoculated with T.001) compare to PS1319 spontaneous fer. undesirable taste of bitter and hearty the overall acceptance showed that it was not significantly different flavor. H. Inoculated fermentations resulted in a chocolate with higher values of resulted as being significantly different (p < 0. F. I) and in the beans (B.3 on the 9-point edonic scale). C. L) during spontaneous and inoculated fermentations.2). overall acceptance of the 5 chocolates. delbrueckii IC103 and T. cerevisiae and T. cerevisiae and T. delbrueckii was the most appreciated PS1319 inoculated with S. SJ02 inoculated with T. PS1319 inoculated with S. 4. 2. delbrueckii and 2I. (2014) recently described how bitter. of undesirable bitter taste and less acid then the inoculated one. coffee flavor. in Bahia. C and D refer to PS1319 conducted with T. alcohol (ethanol) and organic acids (acetic. delbrueckii ID103. Fig. sweet and coffee flavor. with a lower frequency and the volatile compounds. PS1319 the main parameters used to describe the chocolates were desirable and SJ02. as recently reviewed by 37 . was described as fruitier and more roasted than the strates (carbohydrates and citric acid) influence the microbial profile control. delbrueckii monoculture different (p < 0. however duction in astringent. from SJ02 control (scored 7. delbreuckii (scored 7) resulted significantly more acceptable by the post-hoc analysis (p < 0. Answers to the CATA questions highlighted that on fermentation processes conducted on two different hybrids. panel to describe the sensory profile of chocolate was significantly cerevisiae IC67 and T. as diverse hybrids with different characteristics and different initial sub- reported in Fig. Data are expressed as concentration of g/kg. Visintin et al.001) by the Friedman desirable bitter taste. E and F mean PS1319 control. E. delbrueckii ID103. 2 L to SJ02 spontaneous fermentations. woody. Data of the shows CATA results for chocolate profiles of PS1319 fermentations. test. D. on the contrary SJ02 was more sweet. produced from spontaneous and inoculated fermentation. lactic and citric acids) detected in the pulp (A. G. 3B. International Journal of Food Microbiology 257 (2017) 31–40 Fig.

1 ± 0.6 ± 0.8 ± 0.3 ± 0 0.1 0. pungent.1 NS Palmitinic acid Acid 0.3b 2.1 0. delbrueckii have been commercialized and are available to the Indeed.1a 0. sweet.7 ± 0.1 ± 0 NS ⁎⁎ Phenethyl acetate Esters Rose. sweet.1 ± 0.1 NS ketones ⁎ Furfural Aldehydes and Almonds 0.2 Pyran-4-one ⁎ ⁎ Total 6.3 ± 0 NS 0.05). cheese.1 NS ketones a b b ⁎⁎ 2-pyrrolidinone Aldehydes and 0.2 16.2 NS 1-octanoic acid Acid Sweet. rancid 1.2a 8.2 0.2 ± 0 0.1 ± 0 0. IC103 strain to be selected as suitable cocoa beans fermentation starter The yeast metabolism results in ethanol production.6 ± 0.2 NS 4-methylvaleric acid Acid nd nd 0.1 3.1 ± 0a 0.1b 0.1 ± 0.2 NS ⁎ ⁎ 2.6 ± 0.2 ± 0.9 ± 0b 0.3a 11.9 ± 0.2b 41.3 ± 1.7 ± 0.5 ± 0.2 7.1a 2.2 ± 1.1 ± 0.2 NS 1.7 11.8 ± 1.3 ± 0.1b 3.1 0.2 2 ± 0.8 ± 0c 1.6 3.1 0. Furthermore.3 ± 0 NS ⁎⁎ 2.3-butanediol Alcohols Cocoa butter 49.8 ± 0c 46.8a 65.3 NS ⁎⁎ Gamma-butyrolactone Lactones Smoked 0.2 1.1 0. T.5 ± 0.2c 3.7 ± 0.4 ± 0.1 ± 0.8 ± 1.3 ± 0.4 ± 0a 0.8 ± 0c 1.3 ± 0b 0.1 ± 0 1.2 NS ⁎ Isovaleric acid Acid Sweat. Data are expressed as relative percentage from total area of the peaks. some strains 2016) that may have an important impact on chocolate aroma profile.4 ± 0 NS 1. fragrant.1 1 ± 0.5 ± 0.05.1 ± 0b 5.1 ± 0a 0.9 ± 0.2 NS Total 0.2 0. respectively.1 1.3 NS ⁎⁎ Total 0.3 ± 0b 0.2 ± 0b 0.4 ± 0.5a 1.1 NS ketones a b a ⁎⁎⁎ Total 8.6 ± 0.6 ± 0 0.1 NS 0.1 ± 0 0.7 ± 0.2 ± 0 nd nd nd nd Methyl palmitate (palmitic acid. delbrueckii winemaking industry.4 2.2 ± 0 NS ketones roasted ⁎ 2-pyrrolaldehyde Aldehydes and 0.1 ± 0.1 ± 0a 0.1a 0.9 ± 0.3 ± 0.5 ± 0.7 ± 0.3-dihydro-benzofuran Furans 0. grass 0.8ab 19 ± 1.4 9.4b 2.7 ± 0.1c 1 ± 0b 0.2-propanediol Alcohols 1.3 ± 0.5 ± 0.7 ± 0.7 ± 0.1 ± 0.8 ± 0. delbrueckii was not regularly found.6 ± 0. Concentration (%) Compound Group Odor attribute Sample code PS1319 ST PS1319 T PS1319 Sig. 0.1 NS 0.7 ± 0a 0. honey 6.1 ± 0a 0.6 ± 0. 2-phenylacetic acid Acid Floral.2 4.3 ± 0.3 ± 0.1b 1.4 ± 0.3 ± 0 0.4 ± 0 NS Ethylphenyl acetate Esters Fruit. delbrueckii ID103.2 ± 0.1 NS 0.1 0.2 ± 0 0.3a 11. In sample code ST means inoculated with S.7 3.3a 4.7 ± 0.2b 1. 1998).2 1.7 ± 0. ⁎⁎.1b 0.1b 3. microbial activities consume the citric acid 38 .6 ± 0.1a 0. floral 1.2 ± 0. and T with T. of T.9 ± 0.4 ± 0. honey 0. IC103 positively perceived as contribute to the flavor of alcoholic beverages strain was characterized for its enzymatic activities (Visintin et al. Visintin et al.2a 8.7 ± 3. Ciani and Maccarelli.4 ± 0 0.6 ± 1.8 47. honey.8 ± 0.3 NS 0.5 ± 0 0.4 ± 0.9 ± 0.3 ± 0 0.2 0.1 NS Ethyl linoleate (òinoleic acid ethyl ester) Esters 0.3 ± 0.7 3.8 ± 0.2 19.5-dihydroxy-6-methyl-4H.. rancid 3.1 ± 0b 5.4 6.6 NS floral ⁎ Benzyl alcohol Alcohols Sweet.1a 1.4 4. which has been aroma.7 ± 0.6 ± 0a 0.4 ± 0.3 ± 0.3 1.2 ± 0b 0. cerevisiae ID67 and T.2b 0.1 NS ⁎ Hexanoic acid Acid Sweet. Pyrans 6.1 ± 0.3 ± 0.1 NS 1. SJ02 T SJ02 Sig. Different letters a–c within fer- mentations of the same hybrids indicate that are significantly different (p < 0.8 ± 0b 0.2 ± 0b 0. Schwan and Fleet (2014).3 ± 0.2 1.1 ± 0.1 1. fatty 0.3 ± 0 NS ester) b b a ⁎⁎ Total 7.6 ± 0.2 ± 3b 35.1a 1.2 ± 0.6 ± 0. fatty 0.8 ± 0.1 0.1 ± 1.3 ± 0.2 ± 1 NS Ethyl oleate (oleic acid ethyl ester) Esters 0.6a 20. rancid 0.4 2.8 ± 0.1 2 ± 0.2 NS ⁎ Gamma-crotonolactone Lactones 0. International Journal of Food Microbiology 257 (2017) 31–40 Table 3 Concentration of volatile compounds obtained by GC–MS analysis of chocolate produced from spontaneous and inoculated PS1319 and SJ02 hybrids fermentations.3 ± 0.9 ± 0.2a 3.3 0.1 NS 0.7 ± 0 1-nonanoic acid Acid Green.3 8. Moreover.5 NS ketones Benzaldehyde Aldehydes and Bitter almonds.1 1.4 NS 10 ± 0.2b 2.9 ± 0.7 ± 0b 0.6 ± 0. faecal.8 ± 0.3 ± 0.3-dihydro-3.1 NS 0.4 ± 0.2 ± 0.9 ± 0. 2008.2b 13.1a 61.2 ± 0. sweet. esterase. which will be culture. 0. T.2 ± 0 0.1 ± 0 nd nd nd nd rancid Propanonic acid Acid Pungent.01 and p < 0.2 ± 0b 0.8 ± 0.8 ± 1. p < 0.7 ± 0.8 NS ⁎⁎⁎ 2-phenylethanol Alcohols Rose.1 0.3 ± 0.2 8.4 ± 0.9 5.8 ± 0a 1.7 ± 0 nd ± ± ⁎⁎ Total 2.4 NS ⁎⁎ Ethyl dodecanoate Esters Fruity.4 ± 0.4 ± 0.3-Benzopyrrole (indole) Pyrrols 0. delbrueckii has been frequently described as an important transformed to acetic acid by AAB increasing the temperature during yeast in wine sector showing high fermentation ability with production the fermentation with a consequent perceptible strong vinegar-like of very low levels of volatile acidity and acetaldehyde.9 ± 0.3 ± 0.7 ± 0.2 2.7 ± 0.1 0.6 ± 0. sweaty.1 NS ⁎⁎ Total 58. (Bely et al.2 ± 0 NS ⁎ Furfuryl alcohol Alcohols Cooked-sugar 0. oily.1 ± 0 0.8 ± 0.1 0.6 ± 0.7 ± 0.S.2 53..6 NS ⁎⁎ 2.7 8.4 0.1a 3.1 0.2 NS Statistical data are expressed as ⁎.2 0.2 ± 0 NS 1-decanoic acid Acid Rancid.3 ± 0 NS ⁎ Ethyl myristate (Myristic acid ethyl ester) Esters 0.1 NS ketones a b a ⁎⁎ Acetoin Aldehydes and Butter cream 3.4 ± 0 0.9 ± 0b 51.2 1.3 1 ± 0.5 ± 0 0.1 ± 0.6 ± 0.7 ± 0.3 1 ± 0.2 7.4 ± 1.9 ± 1.4 ± 0. roasted-nuts 0.4 ± 0.2 NS ⁎ 2-acetylpyrrole Aldehydes and Chocolate.6 ± 0. honey.001 and no significance.9 ± 0.4 ± 3.2 ± 0.7 ± 0. delbrueckii ID103. flower 0.2 0.1ab 0.6 ± 0.1a 1.2 ± 0.5 ± 0.methyl-4-pyrone (Maltol) Acid Malt.4 NS 1-pentanoic acid (valeric acid) Acid Putrid. ⁎⁎⁎ and NS that indicate significance at p < 0. acid.6 ± 0.5 ± 0.2 ± 0.1 0.4 ± 0 0.2b 0.6 3.5 ± 0 1.1 NS 0. methyl Esters 0.5 5. fatty 1.1 NS 0.5 ± 1 12.3 ± 0 0.6 ± 2.1a 0.2 ± 1.6 ± 0.6 ± 0.1a 2.1 1.7 ± 0.2 ± 0a 0.1 ± 0.1a 1.1 ± 0. cocoa.6 ± 0.3 ± 1.3 ± 0 0. β-glucosidase and glicosidase made T.4 ± 0 0. hazelnut 2.1 NS 1.7 ± 1.9 ± 0.7 NS ketones ⁎⁎ 2-phenyl-2-butenal Aldehydes and flowery.4b 2 ± 0.1 ± 0.7 ± 0.4 ± 0 ⁎ Solerone Lactones Fruity 1.1 NS ⁎ Total 15.8 ± 0.1 0.2 ± 0 0. The values are the mean ± standard deviation of two determinations. flowery 4.2 0.1 NS 0.1 0.8 ± 0.3 ± 0c 0.4 1. 7 ± 0.2 NS ⁎ ⁎⁎ 3-hydroxy-2.2 7.7 ± 0.7 ± 0.

the type of ceptable chocolates with no differences in sensory rankings. the cocoa flavor perceived studies (Batista et al. The center of the diagram corresponds to the lowest flavor intensity and the perimeter to the highest flavor intensity. cerevisiae ID67 and T. 1-nonanoic and 1-dodecanoic acids in chocolate samples. inoculated with T.. the control showed a faster the contrary this last was described as more sweet. within inoculated and spontaneous SJ02 fermentation chocolates. 2-phenylethanol. Visintin et al. since clusters were composed by strains the S. drying and roasting processes. The counts pyrazines. The use of qPCR for complex matrix has ception of the chocolates appeared to exist. hexanoic. del. highest concentration of 2-acetylpyrrole. Moreover. phenethyl acetate and 2- evisiae. As likely. For slower temperature raise compare to the spontaneous processes. took over phenylethanol compounds. Zott et al. 3. however some assays showed greater and/or faster periment. plex matrix. a precise link between volatile compounds and specific and this specific approach has already been previously used in other sensory attributes is difficult. 2015. changes than others. Ethanol produced by S. Within these com- chemical reactions and chemical changes within the bean itself. cere. 2016). cerevisiae and T. the most strongly fermenting ethanol-tolerant species. acetylpyrrole and acetoin (butanoate metabolism). delbrueckii inoculation seemed Some of the herein detected volatile compounds have been reported to accelerate the temperature and pH increase for the hybrid PS1319. 1- indeed AAB reached higher concentration faster than other PS1319 octanoic. as responsible for producing desirable note flavors and off flavors in PS1319 and SJ02 hybrids inoculated with T. S. few differences have been observed by CATA test. Quantitative PCR underlined a higher amount of S. likely aroma compounds as 2-phenylacetic acid. 2008). 2- ture and pH) that killed the beans. PS1319 T and SJ02 T) and cocoa beans spontaneously fermented (PS1319 and SJ02). is in agreement with overall acceptance of sensory analysis. 2012). showing similar shell weights and ac.. during PS1319 mixed ex- present study. delbrueckii than in others. wherein a quency of undesirable bitter taste and less acid than the inoculated one. 2015. whereas mentation inoculated with S. A good correlation between the volatile aroma profile of inoculated fects that may have caused a delay in cycle measurement and thereof an and spontaneous PS1319 and SJ02 fermentations and the sensory per- underestimation of the counts. PS1319. examples. which corresponds well with brueckii expected behavior. delbrueckii ID103. believed to be relevant for the cho- may not be necessary for a successful cocoa fermentation. Fruity and roast In agreement with a slower fermenting ability of T. as expected. with a lower fre- consumption in sugars with respect to the inoculated one. International Journal of Food Microbiology 257 (2017) 31–40 Fig. 2-phenylacetic acid and acetoin. present in the pulp causing an increase in pH. carbohydrates (glucose and fructose) No significant differences in the volatile compounds were noticed and citric acid consumption were observed at the initial times of fer. The presence of pyrazines can be (2015) reported that beans fermented in the presence or absence of LAB influenced by different parameters such as the type and time of fer- were equally fully fermented. cerevisiae and T. contrary. delbrueckii during spontaneous fermentation confirmed the atti. 39 . the weather condition. Temperature and pH delay in the acetic acid production was also probably correlated. and T with T. Aroma profiles of the chocolates produced from PS1319 (A) and SJ02 (B) cocoa beans fermented with inoculations of yeasts starter culture (PS1319 ST. that are pounds some are well known coming from yeast metabolism such as essential for the development of the complex flavor of chocolate phenethyl acetate.. thereby causing an array of bio.. alcohols such as 2. On the visiae and T. cer. delbreueckii. at the descriptors have been used with more frequency respect the control. delbrueckii inoculated chocolate corre- of T. (Pereira et al. the use of starter cultures as being fruity. by GC–MS with respect to the chocolate made by cocoa beans from Results of (GTG)5-PCR fingerprinting allowed us to evaluate the im. delbrueckii seemed to have influenced AAB fermentation. on beginning of SJ02 fermentations (24 h). cocoa. Quantification data by the tasters in the PS1319 T. Particularly. GC–MS also found small amounts of propanoic. the fermentation at the second and the fifth day. plantation of S. Cluster analysis results were in agreement with T. In changes were detected for all different fermentations performed in the agreement with a higher amount of cells. S. In the present study LAB were not con. Since chocolate is a com- already been validated (Schwendimann et al. which may have a negative effect on cocoa aromatic quality.S. colate flavor (Afoakwa et al. On the contrary. trials performed in this study. To note sidered due to the fact that recently it has been hypothesized that LAB the lack of detection of pyrazines. 2004).. delbrueckii ID103. the chocolate processing can influence the development of delbrueckii populations in the inoculated boxes. delbrueckii used as starter culture at Judges described the chocolates made from beans fermented with strain levels. delbreuckii inoculated chocolate having the isolated during the first 24 h of fermentation. mentation. cerevisiae and T. Menezes et al. Nevertheless.. determined might be even higher if we consider a possible matrix ef. cerevisiae and T.3-butanediol and 2-phenylethanol have Inoculated cocoa bean fermentations were characterized by higher been reported in cacao fermentation and these compounds are desirable counts of yeasts during the initial phase. delbrueckii. chocolate profile was the one that showed highest values of Microbial activity generated metabolites and conditions (tempera. delbrueckii showed a cocoa beans during the fermentation. 2010). lated well with the high concentrations of Strecker aldehydes measured tude of this species of not being so commonly involved in this process. furfuryl alcohol. cerevisiae and T. this mentation (until around 72 h) and it was faster in the PS1319 fer. Ho et al. Fermentation with ST code means inoculated with S. an increased amount of acetic acid was recorded. spontaneous fermentations. for high quality cocoa products (Schwan and Wheals. the ripeness of pod. variations in pulp/bean percentage and the storage of pods.

A. with par- starter cultures were able to grow and succeed during cocoa hybrids ticular reference to the impact of fermentation.E. Allegaert. Environ. 250–258. de Andrade e Silva. L. R.S. 1993.. Urso. Vrancken. 14. M. Afoakwa. D. K.. More than 600 volatile compounds are reported behavior of wild yeast during spontaneous wine fermentation in steel tanks and to mark up the complex mixture that characterizes chocolate aroma amphorae. Lupski.. A.. Microbial ecology of artisanal products from North West of Italy and anti- et al. Rodriguez-Campos. 1998. 205–221.. Bisson.G. Int. cerevisiae UFLA CA11 inoculation on mi- crobial communities and volatile compounds of cocoa fermentation.. cerevisiae strain. Z.. Orozco-Avila. J. 205.. Flavor formation and character Jaramillo-Flores. K. Pichia kluyveri and Hanseniaspora uvarum during Schwan. 72–87. Dal Bello. H.. 2014. Kauf. Civera. Dubourdieu. 732–746. Food Res. T. H. Paterson. 2002). Ramos..B. Claisse. and modify some of Microbiol. J.. Nielsen. L.. 2011. cerevisiae. Characterization of the yeast ecosystem in grape must and wine using real-time PCR. Study of the acid bacteria isolated from heap and box cocoa bean fermentations in West Africa. 30. 2011. 2016. V. J.. FEMS Yeast Res. M.. The impact of chocolate quality. G. D. LWT Food Sci. Mills. G.F. International Journal of Food Microbiology 257 (2017) 31–40 The results showed that the consumers were able to differentiate the samples before and after conching. J.. Masneuf-Pomarède. J. Ouwerx.. delbrueckii ID103 “The food of the Gods”: quality determinants of commercial cocoa beans.. C. V.J. FEMS Microbiol.S. cerevisiae yeasts starter used Pereira. P. Use of gas chromatography-ol.N. Appl... Sensory Evaluation Practices. 83–90. L. Teniola. improve the complexity. 840–857. M. Development of a quantitative PCR assay for rapid detection 53. G. R. Pinheiro.a..C. Dias.. Ho et al.. 25–40. L. R. 2016.. L. A.F. 2015. Technol. as also observed by Menezes et al. Schneider. Ramos.... Collin. Counet. 2008. yeast starter cultures on the microbial communities and volatile compounds in cocoa Schwendimann.P. 5395–5405. Paulus. Callemien. The evaluation of T. Investigation of chocolate produced from four different lates. Lett. Schwan yeasts involved in fermentation of West African cocoa beans. J.M... 441–453.. Falony.F.. Alessandria. Menezes.F. Lima. 2008... J.. 2005. Fleet. Wheals. Food Microbiol. fermentations. Fleet.V.. A. 2014. Fischer. 2009. 1883–1894. they influenced the PS1319 hybrid end-pro. 2016. Rev. Rantsiou. P. E. Stoeckle. M. N. J. 44. Pinheiro.. Zwietering. esters and ketones produced by yeasts.E.. Zhao. Food Sci. Eskildsen. stages are applied by the manufactures (Afoakwa et al. Mandarino. L. Saels. Food Sci. Species diversity.S. L. H. S. delbrueckii had a positive influence on the analytical profile of choco. E...F.) during fermentation and drying processes using principal Batista.. 70. V. Yeasts are essential for cocoa bean fermentation. Nielsen. Appl.. Int. G.A... Mathawan. D. 2014.. T. Bellio. K..R. Ramos. Schwenninger. M. Cocoa and Coffee Fermentations.) and S.T. E. L. J. Almeida. C. 2012.... M. R. 5. V. 94–99. M. Amores. acteristics of chocolate. Steensels. Biol. successful changes in aroma profile with respect to spontaneous fer. sociated with wine-making.. This study shows that S. Daniel. Z..L... Int. no sul da Bahia. Owusu.C. Microbiol.. J. 2011.. Group). P. Cocolin. Rantsiou. Food Sci.. Ramos. De Bruijn. Lugo-Cervantes. 306–316..O.M. 2000. Environ... Molecular identifi- fermentations. Food Res.J. A. Crit. 122... P... L. 27. how the steps of fer. R. C. but this may Meersman.H.. Saerens... S. H. As well. Int... Petersen. Swiegers.. Mol. differences observed were too small to significantly change consumer Nielsen. Sidel.. C.M.B. Occurrence and diversity of fermentative process (Afoakwa et al.. 2016.. 54–67. D. enhance the quality.. Microbiol. 43. Escalona-Buendía. and Wheals. G.T. G. Int. and certain cocoa varieties but not for all.. P. T. R. Food Res.. Gantenbein-Demarchi.. 2385–2391. C.. delbrueckii and S. ecology of fresh sausages and characterization of populations of lactic acid bacteria Int.F. M. 81–87. and metabolite kinetics mentation (Batista et al. Schwan. in some cases the cerevisiae. underlining the importance of these microbial species for a suc- (2016). Int. Lefeber. N.. L.G.C. Tuning chocolate flavor through development be improved by using a higher starter inoculations or a co-culture with of thermotolerant Saccharomyces cerevisiae starter cultures with increased acetate S. Food Sci... F. Recently studies using yeasts starter cultures report Papalexandratou. by molecular methods fermented sausage. Schwan. fermentation is considered by the literature to be microbial activity of the autochthonous populations.S. The different results between SJ02 and PS1319 experiments suggest Papalexandratou. H. guards liberation of flavor precursors that are necessary for the for- Ho. The microbiology of Ghanaian cocoa fermentations analysed using culture- perception of the chocolates.N. R.. Appl. 2016.D. Rev.. 81.C. Schwan. that the use of these starter culture strains may be appropriated for Hanseniaspora opuntiae.. Coulon. Fieseler. Ambrosoli. 1101–1110. G..D.. D. (Counet et al. Holzapfel. I. Biomed. A. Biotechnol. O. Ribeiro.. Oenological properties of non-Saccharomyces yeasts as. Molina..H. factometry to identify key odorant compounds in dark chocolate.M. A.. Food Microbiol.R.. Microbiol. Dias. 199–203.. T. Jimenez. Valente... T. 168–186. Lima et al.R. Agric. Fowler.P.. cerevisiae ID67 and T. 2004. Efraim. Romanens. A. Flavor characteristic of chocolate is determined by the ge.. O. 35. Technol.. Larsen. Comparison of Food Microbiol. Technol.G. Res. Direct profiling of the yeast dynamics in wine Visintin. 189..L. 2004. LWT Food Sci.. 2014).. Food Res.. 2013.F. Dynamic components analysis. 2010. 2008.. Daniel. Schwan.J... Impact of netically potential of the cocoa bean variety. I. cessful cocoa. Int. 908–918. Theobroma cacao L. Sena Gomes. cohols. Bahrim...N. No big influenced were recorded for SJ02 hybrid.. In particular. Andersson. Dynamics of volatile and non-volatile compounds in in cocoa and chocolate: a critical review.. Struyf. 40 . 13. B. the most important factor influencing the flavor quality of cocoa and 1151–1159. Meersman et al.. 2013. of Lactobacillus plantarum and Lactobacillus fermentum in cocoa bean fermentation. Characterization and dynamic thereof on chocolate. Skovmand- samples..S. Visintin et al. parameters of the final end-product.D. J. A. Díaz. G. D.T. A. Cocolin. Zhao...C. Miguel. Nout. J.H.. G. Schwan. fermentation. M. D. World J. Zeppa. Masneuf-Pomarede. community dynamics. Microbiol.D. Crit. 2013.. Cocolin.. I. 2007. in combination with the S. behavior of Saccharomyces cerevisiae. 7698–7714. 77. Acetobacter pasteurianus predominate during well-performed Malaysian cocoa bean box fermentations. Microbiological and in the present study might be of interesting value for the cocoa world to physicochemical characterization of small-scale cocoa fermentations and screening of yeast and bacterial strains to develop a defined starter culture. D. LAB and AAB during the Jespersen.F.L.L. Nutr. Miguel. M. 2014. 221–227... Petersen. J. References 64. F. 114. H. 2004). N. S. 2011. Methods 115.M. Int. De Vuyst. 63. The application of starter cultures did change the aroma profile of Brazilian varieties of cocoa (Theobroma cacao L. S. D.. A. Nutr. al. J. Appl. De Vuyst. 69–78.. Maccarelli. Environ. 2015. Nähring. Produtividade do cacaueiro (Theobroma cacau L.. Brasil... Iacumin. 48. Keul.T. Environ. C..... 216. L.R. J.) duct quality changing analytic profile and sensory perception of cho.. ester production. M.. J. Crafack 2010. C. R. Schwan. Academic Press.. Versalovic.. Ryan. C. Stone. 312–320. Jakobsen. Cantoni... Torulaspora delbrueckii-Saccharomyces cerevisiae culture on high-sugar fermentation.. Food Chem. Nutr. An appropriately conducted fermentation safe. bacteria using repetitive sequence-based polymerase chain reaction.L. J.L. W. 2015. Food Microbiol..E. and how the chocolate processing profile of chocolate. J. lactic acid bacteria and acetic Cocolin.. R. Food Microbiol. Food Microbiol.. but also influence the value of acids. L. L.. 78..S. Saccharomyces cerevisiae.. Heimdal. B. M.. 63... L. Ramos. Batista. R. (Taylor & Francis spontaneous and inoculated cocoa fermentations and their effect on sensory char. C. 2008.. impact on those.S. 2014). F. Ho. of the microbiota associated with traditional ecuadorian spontaneous cocoa bean fermentations. The effect of lactic acid bacteria on cocoa bean mation of aroma compounds. M. Fleet. Cell. J. 44. Lee. Pinheiro. Lucas. A. M. Int... Genomic fingerprinting of Ciani.C. C. 1994. cocoa (Theobroma cacao L. E... Hønholt. 51. cultivados em blocos monoclonais. A. Miescher fermentation and the resulting sensory attributes of chocolate. Bely. 73–85. Lonvaud-Funel. N. Food Microbiol. Blennow. E. cation and physiological characterization of yeasts...T. 559–567. Rev. M. A. 2016. Verstrepen.. 50. Ban-Koffi.H. Zott. 82. Crafack. Impact of mixed Microbiol. Comi. C. 2002. 731–761.M. Dolci.... Vol. Moreover. starter cultures and fermentation techniques on the volatile aroma and sensory mentation and drying are carry out. Impact of different cocoa hybrids (Theobroma cacao L. O. S.. K.H. colates. The microbiology of cocoa fermentation and its role in Batista. but in other cases they had a significantly dependent and culture-independent methods. 174. Lactobacillus fermentum.. Crit..R..) inoculated with Saccharomyces the resulting chocolate as determined by GC–MS.. 5. M. California.