Preparative Biochemistry & Biotechnology

ISSN: 1082-6068 (Print) 1532-2297 (Online) Journal homepage: https://www.tandfonline.com/loi/lpbb20

Starter consortia for on-farm cocoa fermentation

and their quality attributes

Yallappa Basappa Saunshi, Mudrakola Vidya Sagar Sandhya, Navin Kumar

Rastogi & Pushpa Srinivas Murthy

To cite this article: Yallappa Basappa Saunshi, Mudrakola Vidya Sagar Sandhya, Navin Kumar
Rastogi & Pushpa Srinivas Murthy (2020) Starter consortia for on-farm cocoa fermentation
and their quality attributes, Preparative Biochemistry & Biotechnology, 50:3, 272-280, DOI:
10.1080/10826068.2019.1689508

To link to this article: https://doi.org/10.1080/10826068.2019.1689508

Published online: 14 Nov 2019.

Submit your article to this journal

Article views: 74

View related articles

View Crossmark data

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=lpbb20
PREPARATIVE BIOCHEMISTRY & BIOTECHNOLOGY
2020, VOL. 50, NO. 3, 272–280
https://doi.org/10.1080/10826068.2019.1689508

Starter consortia for on-farm cocoa fermentation and their quality attributes
Yallappa Basappa Saunshia
, Mudrakola Vidya Sagar Sandhyaa
, Navin Kumar Rastogib , and
Pushpa Srinivas Murthya
a
Department of Spices and Flavour Sciences, CSIR-Central Food Technological Research Institute, Mysore, India; bFood Engineering
Department, CSIR-Central Food Technological Research Institute, Mysore, India

ABSTRACT KEYWORDS
A starter consortium of Saccharomyces cerevisiae (Y), Lactobacillus plantarum (LAB), and Acetobacter CCRD; cocoa; Lactobacillus
aceti (AAB) was defined to ferment the Cocoa beans (Theobroma cacao). Emphasis was laid to plantarum; optimization;
optimize the microbial concentration with a functional ratio of selected cultures. A central com- on-farm fermentation;
starter cultures
posite rotatable design (CCRD) was employed to study the effect of inoculum size (0–20% w/v)
with alcohol, titrable acidity, polyphenols, anthocyanin, cut test, and sensory as response variables.
The significant (p < 0.05) response surface models with high coefficients of determination values
(R2) ranging from 0.82 to 0.93 were considered for the experimental data, which represented the
polynomial response models for describing the constraints. Based on the design, the concentration
of consortia ranged 9.03X103 CFU/g of Y, 5.9X104 CFU/g of LAB, and 7.0X104 CFU/g of AAB. The
graphical optimization of superimposed contour plots fulfilled the desired metabolites;
alcohol (Y1)  11 mg/g, titrable acidity (Y2)  0.25%, polyphenols (Y3)  4.0 mg/g, anthocyanin
(Y4)  14 mg/g, sensory (Y5)  6.0, and cut test (Y6)95%. Thus, validation through a field trial was
confirmed to adopt the techno-economic feasibility on-farm process with precise inoculums. The
effect of starter consortia on Cocoa fermentation and quality was found to be significant.

Introduction The natural cocoa fermentation often leads to the incon-

Raw cocoa beans have an astringent, unpleasant taste and
flavor, and have to be microbially fermented, dried, and
roasted to obtain the desired flavor. The Chocolate quality is
reliant on post-harvest process, which is a prime phase of
cocoa bioprocessing.[1,2] The cocoa pods are cracked to
eliminate the pulp mass and thereby initiates the fermenta-
tion through successive microbial activity of wide range of
yeasts, lactic acid bacteria (LAB), acetic acid bacteria (AAB),
and release a range of metabolic end products which are
also flavor precursors of fermented beans.[1] Saccharomyces
cerevisiae is the most opportunistic species during cocoa fer-
mentation due to its rapid growth, pectinolytic activity and
ethanol tolerance.[3–5] The microbial degradation of pectin
layer by yeasts and LAB results in assimilation of citric acid,
which enhance the pH of the fermenting cocoa mass, com-
ply the growth of anaerobic AAB.[4] The AAB metabolizes
the ethanol to acetic acid through an exothermic process
and diffuses into the beans. The diffused acetic acid disinte-
grates the cellular membranes of cocoa beans and enzyme
activity in the cotyledon develops a characteristic flavor and
color of the fully fermented cocoa beans.[6,7] The volatile
short-chain fatty acids are the key metabolite of the cocoa
bean fermentation process.
sistent fermentation and unreliable quality of the product.
Microbial fermentation induces numerous reactions leading
to a profound biochemical modification of the beans and
formation of organoleptic characteristics.[5] The preliminary
experiments on cocoa fermentation using various inoculum
concentration (10–60%) comprising of Y: LAB: AAB was
carried out by Sandhya et al.[2] A proper succession of
microbiota with their functional roles assures a successful
fermentation process with standard flavor profiles. The data
obtained further suggested that mere inoculum percent is
not sufficient and instigated to confine actual ratio of the
starter consortia inoculum comprising the Y: LAB: AAB to
ferment cocoa beans. Thus, the present study aimed at
developing an optimum inoculum of the microbial consortia
using response surface methodology (RSM) obtained from
central composite rotatable design (CCRD) with Alcohol,
titrable acidity, polyphenols, anthocyanins, cut test score,
and sensory score as response variables. Currently in India,
the fermentation of cocoa bean is carried out as a spontan-
eous/natural on-farm processing. It is vital to upgrade post-
harvest process and standardize fermentation process for
commercial use with emphasis on uniform and consistent
quality characteristics.

CONTACT Pushpa S. Murthy pushpa@cftri.res.in Department of Spices and Flavour Sciences, CSIR-Central Food Technological Research Institute, Mysore
570020, India.

These authors contributed equally to this work.
ß 2019 Taylor & Francis Group, LLC

Materials and methods
Raw material and chemicals
Cocoa pods (Theobroma cacao var. Forastero) were obtained
from Puttur, Karnataka, India during the first crop season
of 2013 (May–Aug). The chemicals used in the experiment
were procured from Sigma-Aldrich, India and microbio-
logical media from Hi-Media, Mumbai, India.
Development of microbial consortium for cocoa
fermentation
The microbial cultures namely Saccharomyces cerevisiae
MTCC 173, Lactobacillus plantarum MTCC 5422, and
Acetobacter aceti MTCC 3347 were procured from IMTECH
Chandigarh, India and maintained in the host laboratory.
Yeast was grown in potato dextrose broth (PDB) at
28 ± 2  C. Whereas, LAB and AAB were cultured in de
Man–Rogosa–Sharpe (MRS) medium and Acetobacter broth
(AAB), respectively at 37 ± 2  C temperature. The cells were
harvested by centrifugation (8000g, 20 min, 4  C), washed
with sterile saline (0.85% NaCl solution, w/v) and used
as inoculum.
Bio-processing of cocoa beans
The harvested fresh cocoa fruits were surface sanitized,
manually cut open to separate the beans from the placenta
and fermentation studies were carried out according to
Sandhya et al.[2] The Cocoa pulp was placed in a wooden
box (0.5 m  0.5 m  0.5 m) with 50 kg capacity, lined with
banana leaves and inoculated with predetermined proportion
of Y, LAB, AAB, and incubated for 72 h. The cocoa mass
was thoroughly mixed every 24 h of fermentation and the
samples were collected at the fixed schedule of 0, 24, and
72 h of fermentation for analysis.[2,3,6–8]
Analysis of various parameters of fermented
cocoa beans
Sample preparation
The cocoa bean mass (100 g) was extracted using sterile dis-
tilled water (100 ml) and filtered using 0.45 mm filter unit
(Hi-Media, India). Thus obtained filtrate was utilized for
physical and chemical investigation. The rest of the freshly
fermented beans were evenly spread on the tray (200 ) and
dried at 60  C for 16 h using a cross flow drier (Tray drier
Premium Model PTD – 48E, Land T Ltd., India) until the
moisture was 5 ± 2%.
Analysis of metabolites and cut test
The alcohol production was estimated using potassium
dichromate and was quantified (mg/g) from the standard
curve at 600 nm absorbance using UV–VIS spectrophotom-
eter (GBC model, India).[10,11] The total acids formed during
fermentation was determined by titrating the filtrate with
PREPARATIVE BIOCHEMISTRY & BIOTECHNOLOGY 273
0.1 N NaOH and phenolphthalein as an indicator.[9,10] The
fermented cocoa bean powder was defatted using petroleum
ether (60–80  C) for 8 h.[10,11] The defatted cocoa powder
was air-dried to remove the solvent and the samples (0.2 g)
was extracted with methanolic HCl (80% MeOH containing
1% HCl) under shaking condition for 2 h at 28  C and cen-
trifuged at 1000 rpm for 15 min. The anthocyanin hydrolysis
was carried out as per the method described by Sandhya
et al.[2] and the total anthocyanins (mg/kg) was evaluated as
follows: Total anthocyanins (mg/kg) ¼ TOD  1000/
(AvE530)1 cm/10, where TOD ¼ total optical density,
(AvE530)1 cm ¼ the average extinction coefficient for total
anthocyanin when 1 cm cuvette and 1% (10 mg/ml) stand-
ards are used.
The total polyphenol content (mg/g) of cocoa bean in
methanolic extracts was estimated by Folin-Ciocalteu
method[12] and the results were expressed in epicatechin
equivalents. The cocoa cut test was performed by longitu-
dinally splitting via center of the fermented, dried cocoa
bean (to expose the maximum surface of the cotyledons)
and examined under daylight. The cut test score was defined
as the percentage of brown bean.
Organoleptic profile of the fermented cocoa beans
The dried fermented cocoa beans (natural and starter) were
roasted using pilot cocoa roaster for 30 min, followed by
deshelling. The nibs were ground to obtain cocoa mass/
liquor. Further, the cocoa liquor were evaluated for organo-
leptic profile by the trained panel of cocoa tasters compris-
ing of women (2) and men (4) in the age groups of
25–55 years at M/s Campco Pvt Ltd., Puttur, India. The spe-
cific flavor notes such as acid, fruity, cocoa, bitter, etc. for
the samples were documented and conferred during the ses-
sions. The taste intensity and overall profile were expressed
in 0–10 Hedonic scale.[2,3,13,14]
Experimental design, statistical analysis, and validation
A three-factor CCRD was employed to determine the appro-
priate concentration of yeast, LAB and AAB ranging from
0% to 20% (w/v) for cocoa fermentation. Twenty experi-
ments were performed as per CCRD (Table 1) with Alcohol,
titrable acidity, polyphenols, anthocyanins, sensory score,
and the cut test score were the responses. A second-order
polynomial equation was used to fit the experimental data
given in Table 2.
The model proposed for the response (Yi) is
Yi ¼ a0 þ a1 X1 þ a2 X2 þ a3 X3 þ a11 X112
þ a22 X222 þ a33 X332 þ a12 X1X2 þ a13 X1X3
þ a23X2X3 þ e :::
(1)
where Yi (i ¼ 1–6) is the response to alcohol (Y1), titrable
acidity (Y2), total polyphenols (Y3), total anthocyanin (Y4),
sensory score (Y5) , cut test (Y6), and are reported in Table
2. The a0 is the value of the fitted response at the center
274 Y. B. SAUNSHI ET AL.

Table 1. Experimental design for three-factor CCRD and response in terms.

Alcohol Titrable Polyphenols Anthocyanin Sensory score Cut test
Yeast (%) LAB (%) AA (%) (mg/g) acidity (%) (mg/g) (mg/kg) (hedonic scale) score (%)
Exp No. X1 X2 X3 (1) (2) (3) (4) (5) (6)
1 1 1 1 6.00 0.39 3.70 13.70 4.0 85
2 1 1 1 6.00 0.31 4.56 12.40 4.0 85
3 1 1 1 10.10 0.28 3.55 11.30 5.0 100
4 1 1 1 5.50 0.36 4.07 12.70 4.0 85
5 1 1 1 5.00 0.34 3.01 11.26 4.0 85
6 1 1 1 5.80 0.30 2.45 10.89 4.0 80
7 1 1 1 10.20 0.24 3.73 11.40 6.0 100
8 1 1 1 5.50 0.28 3.25 13.00 4.0 85
9 1.682 0 0 7.90 0.26 4.06 10.90 5.0 90
10 1.682 0 0 10.00 0.27 4.47 11.50 5.0 100
11 0 1.682 0 5.00 0.34 2.80 12.18 4.0 85
12 0 1.682 0 8.90 0.28 3.34 11.97 5.0 95
13 0 0 1.682 3.00 0.35 4.72 14.80 4.0 45
14 0 0 1.682 3.00 0.28 2.94 15.02 4.0 70
15 0 0 0 4.29 0.27 4.10 12.49 4.0 75
16 0 0 0 3.50 0.29 3.99 11.20 4.0 75
17 0 0 0 4.31 0.27 3.87 12.49 4.0 75
18 0 0 0 3.50 0.27 3.83 11.39 4.0 75
19 0 0 0 4.03 0.26 3.82 11.39 4.0 75
20 0 0 0 4.32 0.28 3.86 11.20 4.0 75

Table 2. Variables and their levels for CCRD. was expressed as surface and contour plots to envisage the
Symbols 1.682 1 0 1 1.682 Mean Std. Dev relationship between the response and experimental levels of
Yeast (%) X1 0 4.05 10 15.95 20 10 5.95 every factor and to deduce the optimum concentration
LAB (%) X2 0 4.05 10 15.95 20 10 5.95 of inoculum.
AA (%) X3 0 4.05 10 15.95 20 10 5.95
The pilot-scale cocoa fermentation was carried out using
the optimized starter cultures using the wooden boxes at a
Table 3. Estimated coefficients of the fitted second order polynomial repre- batch size of 50 kg (harvested pods) at M/s Campco Factory
senting the relationship between the response and the process variable. for validation. The starter cultures (Y: LAB: AAB) obtained
Titrable Sensory Cut test after optimization by RSM were used for fermentation. The
Alcohol acidity Polyphenols Anthocyanin score score
(1) (2) (3) (4) (5) (6)
cocoa mass was simulated with starter culture, incubated for
a0 3.98a 0.27a 3.92a 11.72a 4.00a 74.78a
72 h and intermittent mixing was done till the completion of
a1 0.36ns 0.00ns 0.08ns 0.17ns 0.22b 1.33ns fermentation. The assays were conducted for secondary
a2 1.10b 0.02a 0.13b 0.01ns 0.34a 3.79c metabolites as described above.
a3 0.08ns 0.02a 0.47a 0.23ns 0.07ns 2.71ns
a11 1.82a 0.00ns 0.09c 0.32c 0.32a 8.53a
a12 1.26b 0.03a 0.03ns 0.58c 0.37b 3.13ns
a13 0.09ns 0.00ns 0.30a 0.14ns 0.13ns 0.62ns Results and discussion
a22 1.11b 0.02a 0.33a 0.01ns 0.14c 6.77b
a23 0.16ns 0.01ns 0.27a 0.54c 0.13ns 0.63ns The uncontrolled nature of the natural fermentation process,
a33 0.28ns 0.02a 0.06ns 0.99a 0.03ns 4.72c particularly with respect to the lack of control over the
a
b
Significant at 0.1% level. growth, development of microorganisms and metabolic pro-
Significant at 1% level.
c
Significant at 5% level.
duction during the process influences the quality of the fin-
ns
Nonsignificant even at 5% level, n ¼ 3. ished cocoa beans and is also inconsistent. Hence the
regulation of cocoa fermentation with precise starter cultures
point of the design, a1, a2, a3 being the linear, a11, a22, beside inoculums size in ratio of the beans (kg) is essential.
a33 being the quadratic, a12, a13, a23 being the cross- Thus, the cocoa fermentation was conceded with the active
product terms, respectively and e is the random error. The microbes such as Yeast, LAB and AAB selected based on the
coefficients of Eq. (1) were obtained using MATLAB 7.0 respective metabolite production necessary for obtaining the
software (The Math Works Inc., Natick, MA, USA) and are desired product.[2,15] In the fermentation process, the phys-
presented in Table 3. The analysis of variance (ANOVA) ical and chemical factor plays a crucial role for efficient end
was evaluated to determine the significance of regression product formation and hence the optimum inoculum by
coefficients of individual linear, quadratic and interaction using statistical approach for consistent fermentation
terms. The goodness of fit of the model was checked by the was explored.
determination coefficient (R2) values (Table 4).
Maximization of fitted polynomials for the responses such Development of microbial consortium for cocoa
as alcohol content, titrable acidity, total polyphenols, total
fermentation
anthocyanins, cut test score, and sensory score were per-
formed by a non-linear mathematical maximization proced- The microorganisms selected as starter cultures are expected
ure of the Quattro Pro software package 4.0 (Borland to have several characteristics such as easy adaption to the
International Inc., USA). The fitted polynomial equation raw material, developing sensory quality, extending shelf life,
 PREPARATIVE BIOCHEMISTRY & BIOTECHNOLOGY 275

Table 4. ANOVA for the fitted polynomial model as per CCRD.

Sum of squares
Alcohol Titrable Polyphenols Anthocyanin Sensory score Cut test
(mg/g) acidity (%) (mg/g) (mg/kg) (hedonic scale) score (%)
Source of variation df (1) (2) (3) (4) (5) (6)
Regression
First-order terms 3 18.50 0.012 77.42 1.14 2.34 321.31
Second-order terms 6 77.89 0.015 53.99 21.98 3.12 2191.16
Total 9 96.39 0.028 131.41 23.13 5.46 2512.47
Residual
Lack of fit 5 9.85 0.001 7.68 1.73 0.69 400.53
Pure error 5 0.89 0.001 6.99 2.77 0.05 157.00
Total error 10 10.74 0.002 14.67 4.50 0.74 557.53
Grand total 19 107.13 0.030 146.08 27.62 6.20 3070.00
R2 0.90 0.929 0.90 0.84 0.88 0.82

Table 5. Microbial ecology and enumeration (log CFU) in cocoa fermentation.

Fermentation duration (days)
Fermentation Organisms 1 2 3 4 5 6 7
Natural Y 3.1 ± 2.0 6.8 ± 5.5 7.2 ± 6.0 7.1 ± 0.5 7.7 ± 0.1 1.5 ± 0.5 0.8 ± 0.2
LAB 3.0 ± 1.8 6.5 ± 2.1 7.5 ± 1.5 7.2 ± 1.2 7.8 ± 0.8 5.6 ± 0.5 4.4 ± 0.2
AAB 1.9 ± 0.2 4.2 ± 0.8 6.5 ± 0.5 7.4 ± 0.8 8.9 ± 1.0 9.2 ± 1.2 10 ± 2.0
Starter Y 3.9 ± 0.1 8.7 ± 0.8 7.4 ± 0.2 3.4 ± 0.4 – – –
LAB 4.7 ± 0.4 8.8 ± 0.7 7.8 ± 0.5 4.2 ± 0.6 – – –
AAB 4.84 ± 0.8 6.3 ± 0.9 8.5 ± 0.7 9.8 ± 0.5 – – –

Data expressed as mean ± SD, n ¼ 3. Microbes such as Pichia, Bacillus subtilis, Candida species, etc. were observed in Natural fermentation. Starter culture
fermentation was completed on the third day.

minimize the processing time, inhibiting food-related patho- Table 6. Comparison of the quality parameters of the natural and starter fer-
gens and non-toxigenic properties.[19]. The starter strains mented cocoa.
were screened and sourced from pectin rich citrus peel, Parameters Natural Starter
apple, grapes, and coffee pulp. The initiation of the fermen- Moisture (%) 8.7 ± 0.2 8.0 ± 0.1
pH (1–14) 5.7 ± 0.4 4.4 ± 0.4
tation rapidly increased in population, specifically, the yeast Temperature ( C) 41 ± 1.4 42.5 ± 2.5
from 3.9 to 8.7 log CFU/ml, by consumption of mucilage Cut test score (%) 80 ± 0.2 100 ± 0.2
sugar. Eventually, the LAB growth was recorded to increase Alcohol (mg/kg) 14.5 ± 1.1 11.8 ± 1.5
Anthocyanin (mg/kg) 12 ± 1.2 14 ± 1.4
4.7 log CFU/ml to 8.8 log CFU/ml whereas, AAB ascended Acids (%) 0.9 ± 0.1 0.4 ± 0.1
4.8–6.3 log CFU/ml at 24 h. The starter micro-biota utilizes Sugars (%) 1.15 ± 0.2 1
the carbohydrate for growth and regulation of fermentation Polyphenols (mg/g) 2.75 ± 0.5 3.75 ± 0.7
Sensory (1–10) 5.5 ± 0.7 6.75 ± 0.3
process. The naturally fermented cocoa was observed to
Data expressed as mean ± SD, n ¼ 3.
have less microbial population with yeast 6.8 log CFU/ml,
LAB 6.5 log CFU/ml, AAB 4.2 log CFU/ml at 24 h
yeast, LAB and AAB on the cocoa fermentation on the
(Table 5). Similar results were corroborated in the previous
responses namely alcohol (Y1), titrable acidity (Y2), poly-
study reported.[2,20] phenols (Y3), anthocyanin (Y4), sensory score (Y5), and cut
test score (Y6) are represented by the coefficients for the
Physico-chemical composition of fermented cocoa beans actual functional relations of second-order polynomials for
predicting responses (Yi) is presented in Table 3. The insig-
Based on the geographical condition, cocoa variety the nificant terms were excluded based on Student’s t-ratio. The
mucilage composition of the cocoa beans varies. The param- responses to different combinations were analyzed using
eters obtained for starter treated cocoa beans were signifi- ANOVA appropriate to the experimental design (Table 4).
cant compared to natural fermentation. The analyzed It reflected that the sum of squares due to regression (first
physico-chemical composition of fermented cocoa beans and second order terms) was important and the lack of fit
were effectively influenced by starter cultures and relative was not considered at 1% or 5% level. The high values of
inoculums size with improved cut-test score (100%) and the coefficient of determination (R2) also suggest that the
sensory (6.75) compared to naturally fermented (cut-test model fitted well with the experimental data. An admirable
score; 80% and sensory; 5.5) cocoa beans (Table 6). statistical model revealed, the R2 in the range of 0–1.0 and
the value closer to 1.0 signifies a better fit of the model. The
response surfaces were selected based on the observation of
Analytical scrutiny of the models, optimization
the data and initial optimization of the individual responses
and validation
(Fig. 1). The values of second order polynomial equations
Fermentation determines the final quality of products pro- (Eq. (1)) for maximum sensory analysis (Y5) and cut test
duced, especially flavor. The consequence of application of score (Y6) as well as minimum alcohol (Y1), acidity (Y2),
276 Y. B. SAUNSHI ET AL.
Figure 1. Response surfaces presenting the effect of yeast and LAB concentration on (a) alcohol (Y1, mg/g), (b) titrable acidity (Y2, %), (c) polyphenol (Y3, mg/g),
(d) anthocyanin (Y4, mg/kg), (e) sensory score (Y5, hedonic scale), and (f) cut test score (Y6, %).
polyphenols (Y3), and anthocyanin (Y4), established on the
coefficients provided in Table 3 have been presented in
Table 7. The independent variable AAB (X3) (20% coded
value þ1.682) was constant for all the responses. Hence, to
deduce workable optimum conditions, the graphical opti-
mization technique was adopted by considering AAB (X3)
as the programed optimum condition.
The contour plots for the response were generated and
were compared (Fig. 2). The specifications necessary for
each response were first set, which also served as constraints
for optimization. An acceptable compromise was made fol-
lowing the criteria for the alcohol (Y1)  11 mg/g, titrable
acidity (Y2)  0.25%, total polyphenols (Y3)  4.0 mg/g, total
anthocyanin (Y4)  14 mg/g, sensory score (Y5)  6.0 and
cut test score (Y6)  95%.
The contour plots were superimposed and the regions
that satisfied all the constraints were selected as the optimum
conditions. Based on the results a combination (A, B, C, D,
and E) could be selected from the shaded area (Fig. 3, Table
8). Superimposition of contour plots implies that yeast rang-
ing from 2.58% to 5.62%, LAB ranging from 13.65% to
19.18% and AAB (20%) fulfilled the criteria for optimization.
The RSM plot depicts the interaction effect of Yeast and
LAB (Fig. 1). The alcohol, titrable acidity, and polyphenol
were found to be an active function of LAB (Table 3). The
cocoa mass changes due to the development of yeasts and
LAB, which allow the obligate anaerobic AAB to replicate
and in turn oxidize ethanol into acetic acid.[15–17]
Influence of yeast and LAB concentration on alcohol
and total polyphenols
At the lowest level of yeast (0%, coded value 1.682), alco-
hol production was found to intensify from 5.5 to 16 mg/g
 PREPARATIVE BIOCHEMISTRY & BIOTECHNOLOGY 277

Table 7. Optimum conditions for maximum alcohol, titrable acidity, sensory score, cut test score, and minimum polyphenol, anthocyanin.
Alcohol Titrable Polyphenols Anthocyanin Sensory Cut test
(mg/g) acidity (%) (mg/g) (mg/kg) score (0–10) score (%)
Yeast (%) (X1) LAB (%) (X2) AA (%) (X3) (1) (2) (3) (4) (5) (6)
Independent variables Responses
Conditions for minimum alcohol (Y1)
8.86 5.68 20.00 2.58 0.33 2.42 13.46 3.74 66.40
0.19 0.73 1.68 (2.60) (0.35) (2.50) (14.00) (3.90) (67.20)
Conditions for titrable acidity (Y2)
0.00 0.00 20.00 5.82 0.50 1.91 12.65 4.09 96.20
1.68 1.68 1.68 (6.01) (0.45) (1.99) (13.20) (4.05) (96.80)
Conditions for minimum polyphenols (Y3)
20.00 0.00 20.00 12.23 0.33 0.64 10.73 4.70 105.90
1.68 1.68 1.68 (13.50) (0.36) (0.68) (10.91) (4.99) (106.0)
Conditions for minimum anthocyanin (Y4)
0.00 20.00 20.00 17.53 0.20 4.05 12.39 8.05 130.16
1.68 1.68 1.68 (18.05) (0.23) (4.20) (13.05) (8.20) (130.0)
Conditions for maximum sensory score (Y5)
0.00 20.00 20.00 17.53 0.20 4.05 12.39 8.05 130.16
1.68 1.68 1.68 (18.05) (0.23) (4.20) (13.05) (8.20) (130.16)
Conditions for maximum cut test score (Y6)
4.33 18.27 20.00 10.61 0.24 3.65 13.92 6.27 99.92
0.95 1.39 1.68 (10.99) (0.28) (3.88) (14.10) (6.55) (100.00)
The values in the parenthesis for independent variables are the coded values. For the response, the values in the parenthesis are the mean experimental values
based on the mean value of five determinations.

Figure 2. Contour plots showing the effect of Yeast and LAB concentration on (a) alcohol (Y1, mg/g), (b) titrable acidity (Y2, %), (c) polyphenol (Y3, mg/g), (d)
anthocyanin (Y4, mg/kg), (e) sensory score (Y5, hedonic scale), and (f) cut test score (Y6, %).
278 Y. B. SAUNSHI ET AL.

(20%, coded value þ1.682), alcohol was found to decrease

from 16.0 to 8.0 mg/g initially, and then improved to
9.5 mg/g with an raise in yeast from 0% to 20%, whereas
acid and polyphenol were found to elevated and decline,
respectively along with yeast from 0% to 20%.

Influence of yeast and LAB concentration on

anthocyanin, sensory, and cut test score
Fermentation tests uncovered that beans fermented devoid
of yeasts were purplish-violet, partially brown and was def-
icit of characteristic chocolate flavor.[15,16] At the lowest
level of yeast (0%, coded value 1.682), anthocyanin and
cut test score did not change seriously (Fig. 1e). At the high-
est level of yeast (20%, coded value þ1.682), the sensory
score and cut test score were not significant with amplifica-
tion of LAB from 0% to 20%. The anthocyanin intensified
Figure 3. Superimposed contour plots showing the shaded overlapping
area for alcohol (Y1)  11 mg/g, titrable acidity (Y2)  0.25%, polyphenol (9.5–17.5 mg/kg) with an augmentation in LAB from 0% to
(Y3)  4.0 mg/g, anthocyanin (Y4)  14 mg/g, sensory score (Y5)  6.0, and cut 20% (Fig. 1d). The anthocyanin, sensory and cut test index
test score (Y6)  95%. was found to be a strong function of LAB (Table 3). At the
lowest level of the LAB (0%, coded value 1.682), no better
Table 8. Feasible optimum conditions and predicted and experimental value changes was noticed in anthocyanin and the sensory score
of response at optimum conditions. with the proliferation of yeast from 0% to 10%. Beyond this
Conditions value, the anthocyanin and sensory score were found to
A B C D E reduce and rise, respectively. At the highest level of the LAB
Variables (20%, coded value þ1.682), anthocyanin was found to
Yeast (%) 2.58 3.59 4.24 5.62 3.13 expand, and the sensory score was found to decrease with a
1.25 1.08 0.97 0.74 1.15
LAB (%) 16.94 18.12 19.18 16.47 13.65 rise in yeast from 0% to 20%. For all the values of the LAB
1.17 1.37 1.54 1.09 0.61 (0–20%), cut test score was found to decrease to an extent
AA (%) 20.00 20.00 20.00 20.00 20.00 and then again expanded with elaboration of yeast (Fig.
(1.68) (1.68) (1.68) (1.68) (1.68)
Responses 1f).The suitability of the model equation for envisaging the
Alcohol (Y1) (mg/g) 11.10 11.18 11.61 8.01 7.87 optimum response values was verified using recommended
(11.45) (12.48) (12.49) (9.58) (7.65)
Titrable Acid (Y3) (%) 0.23 0.24 0.24 0.25 0.26
optimum conditions as determined by graphical optimiza-
(0.20) (0.22) (0.23) (0.26) (0.27) tion approach. The conditions for the cocoa fermentation
Polyphenol (Y5) (mg/g) 3.90 3.74 3.60 3.58 3.82 (Yeast, LAB, and AAB) are as shown in Table 8 and consid-
(3.70) (3.95) (3.15) (3.33) (3.61)
Anthocyanin (Y6) (mg/kg) 13.32 13.68 13.94 14.16 13.38
ered as optimal and feasible. The experimental values were
(13.60) (13.50) (13.59) (14.03) (13.83) found to concur with the predicted values (Tables 7 and 8)
Sensory score (Y7) (1  10) 6.46 6.45 6.51 5.60 5.62 evince that our model was precise and reliable.
(6.08) (6.49) (6.50) (5.91) (5.65)
Cut test score (Y8) 100.00 100.00 100.00 88.16 87.84
(%) (100) (98.50) (99.15) (90.33) (89.61)
Field trials
The values in the parenthesis for independent variables are the coded values.
For the response, the values in the parenthesis are the mean experimental A validation study to observe the performance of developed
values based on the mean value of five determinations. Conditions A, B, C,
D, E and F have been indicated in Fig. 3. starters under industrial conditions (pilot scale/field trials)
was executed at the M/s Campco farms. The harvested
Cocoa pods (50 ± 5 kg) with the optimum maturity are piled
in a heap and allowed to remain for 2–7 days at room tem-
with acceleration in LAB concentration from 0 to 20%. The
perature. The pod shell (40 ± 2 kg) and the beans (10 ± 5 kg)
polyphenol raised initially with LAB concentration of 16%
were separated. The cocoa beans (wet beans 10 kg) that were
and then declined (Fig. 1c). At the highest level of yeast
fermented with specific starter cultures [Y: LAB: AAB::4.0%
(20%, coded value þ1.682), alcohol and polyphenols were
(w/v): 16.18% (w/v): 20% (w/v)] denoted suitable candidates
found to differ since yeasts produced ethanol from sugars
for spontaneous fermentation of the cocoa pulp and out-
available from cocoa pulp.[18,19] The LAB (10%) was not
competed the indigenous batch of cocoa with respect to fer-
effective as far as the acid release is concerned, but marginal
mentation quality.
rise with a simulation in LAB above 10% was observed. At
The first 2–3 days of fermentation were characterized by
the lowest level of LAB (0%, coded value 1.682), alcohol
the succession growth of the micro-biota (Y, LAB, and
release was found to decrease from 5.5 to 3.5 mg/g initially
AAB) and appeared to be stable on cocoa pulp fermenta-
and then escalated to 12 mg/g with an upsurge in yeast from
tions. The fermentation process, batch size, climatic condi-
0% to 20% (Fig. 1a). At the highest level of LAB
tions, the bioprocess period, etc. which makes

reproducibility of natural fermentation particularly difficult
needs to be reoriented to use of starter culture. To assess
the microbial diversity of the natural and starters for pilot-
scale fermentations, samples were collected throughout the
fermentation period and suggestive differences in the LAB
and AAB cell counts were observed. The micro-biota
(Yeasts, Candida, Pichia, Bacillus subtilis, LAB, etc.) were
very diverse in natural fermentations and become prepotent
at varying periods and temperature of fermentation.[2]
The Starter induced process is vital to obtain a consistent
microbiome so as to function competently to bring in the
desired metabolites in the specific duration of fermentation
(Table 5). The relative population sizes were calculated
based on a representative number of strains in the fermenta-
tion (Table 6). Also, the discharge of acids, sugars, and alco-
hol was in accordance with lab-scale observations and
previous reports,[1–3] when natural fermentation was com-
pared to that of starters. The results obtained by application
of the developed starter cultures in pilot scale (50 kg) fer-
mentation were significant. Further, adaptation of the tech-
nology would be more economical, consistent with
improved quality chocolate.
Conclusions
The present study reports for the first time that the accurate
inoculum of defined mixed starter culture with 9.03  103
CFU/g of yeast, 5.9  104 CFU/g of LAB, and 7.0  104
CFU/g of AAB and their ratio essential to accelerate cocoa
fermentation with functional role and metabolites. The
developed models could predict the alcohol and acid pro-
duction, polyphenol bleaching, anthocyanin hydrolysis, Cut
test score and Sensory analysis that are prime parameters of
fermented cocoa quality. Superimpositions of contour plots
stipulate that Yeast and LAB concentration ranging from
1.68 to þ1.68 coded values were suitable for achieving
quality cocoa fermentation. The insights from this study
insists a need to develop commercial processes with techno-
economic feasibility.
Acknowledgments
The authors gratefully acknowledge Director, CSIR–CFTRI, Mysuru,
India for providing necessary laboratory facilities. The authors also
thank the M/s Campco Pvt. Ltd, Puttur, Karnataka, India for providing
the cocoa samples and undertaking the sensory evaluation.
Disclosure statement
No potential conflict of interest was reported by the authors.
Funding
The authors also thank the Ministry of Food processing Industry, DST,
Govt. of India, Grant no [SERB/MOFPI/0009/2013] for funding
the project.
PREPARATIVE BIOCHEMISTRY & BIOTECHNOLOGY 279
ORCID
Navin Kumar Rastogi http://orcid.org/0000-0003-0790-5841
References
[1] Camu, N.; De Winter, T.; Addo, S. K.; Takrama, J. S.; Bernaert,
H.; De Vuyst, L. Fermentation of Cocoa Beans: influence of
Microbial Activities and Polyphenol Concentrations on the
Flavour of Chocolate. J. Sci. Food Agric. 2008, 88, 2288–2297.
DOI: 10.1002/jsfa.3349.
[2] Sandhya, M. V. S.; Yallappa, B. S.; Varadaraj, M. C.; Puranaik,
J.; Rao, L. J.; Janardhan, P.; Murthy, P. S. Inoculum of the
Starter Consortia and Interactive Metabolic Process in
Enhancing Quality of Cocoa Bean (Theobroma cacao)
Fermentation. LWT-Food Sci. Technol. 2016, 65, 731–738. DOI:
10.1016/j.lwt.2015.09.002.
[3] Saunshia, Y.; Sandhya, M. V. S.; Lingamallu, J. M. R.; Padela, J.;
Murthy, P. Improved Fermentation of Cocoa Beans with
Enhanced Aroma Profiles. Food Biotechnol. 2018, 32, 257–272.
DOI: 10.1080/08905436.2018.1519444.
[4] Mota-Gutierrez, J.; Botta, C.; Ferrocino, I.; Giordano, M.;
Bertolino, M.; Dolci, P.; Cannoni, M.; Cocolin, L. Dynamics
and Biodiversity of Bacterial and Yeast Communities during
Fermentation of Cocoa Beans. Appl. Environ. Microbiol. 2018,
84, e01164–18.
[5] Ferreira, R.; Teixeira, P. G.; Gossing, M.; David, F.; Siewers, V.;
Nielsen, J. Metabolic Engineering of Saccharomyces cerevisiae
for Overproduction of Triacylglycerols. Metab. Eng. Commun.
2018, 6, 22–27. DOI: 10.1016/j.meteno.2018.01.002.
[6] Nielsen, D. S.; Teniola, O. D.; Ban-Koffi, L.; Owusu, M.;
Andersson, T. S.; Holzapfel, W. H. The Microbiology of
Ghanaian Cocoa Fermentations Analysed Using Culture-
Dependent and Culture-Independent Methods. Int. J. Food
Microbiol. 2007, 114, 168–186. DOI: 10.1016/j.ijfoodmicro.2006.
09.010.
[7] Thompson, S. S.; Miller, K. B.; Lopez, A. S.; Camu, N. Cocoa
and Coffee. In M. P. Doyle, L. R. Beuchat & T. J. Montville
(Eds.), Food microbiology: fundamentals and frontiers, 4,
881–899. Washington: ASM Press.
[8] Carr, J. G.; Davies, P. A.; Dougan, J. Cocoa Fermentation in
Ghana and Malaysia. In 7th International Cocoa Research
Conference, Douala, Cameroon, Cocoa, Chocolate and
Confectionery Alliance of the United Kingdom, London (RU),
1979.
[9] Ardhana, M. M.; Fleet, G. H. The Microbial Ecology of Cocoa
Bean Fermentations in Indonesia. Int. J. Food Microbiol. 2003,
86, 87–99. DOI: 10.1016/s0168-1605(03)00081-3.
[10] Biehl, B.; Voigt, J.; Heinrichs, H.; Senjuk, V.; Bytof, G. pH-
Dependent Enzymatic Formation of Oligopeptides and Amino
acids, the Aroma Precursors in Raw Cocoa Beans. In
Proceedings in XIth International Cocoa Research Conference,
Cocoa Producers’ Alliance, Yamassoukro, Ivory Coast, 1993.
[11] Williams, M.; Reese, D. Colorimetric Determination of Ethyl
Alcohol. Anal. Chem. 1950, 22, 1556. DOI: 10.1021/
ac60048a025.
[12] Nazaruddin, R.; Seng, L. K.; Hassan, O.; Said, M. Effect of Pulp
Preconditioning on the Content of Polyphenols in Cocoa Beans
(Theobroma cacao) during Fermentation. Ind. Crops Prod. 2006,
24, 87–94. DOI: 10.1016/j.indcrop.2006.03.013.
[13] AOAC Official Methods of Analysis. Association of Official
Agricultural Chemists. 18th ed.; AOAC Official Methods of
Analysis: Washington DC, 2005; pp. 14.
[14] Misnawi, J. S.; Jamilah, B.; Nazamid, S. Effects of Incubation
and Polyphenol Oxidase Enrichment on Colour, Fermentation
Index, Procyanidins and Astringency of Unfermented and
Partly Fermented Cocoa Beans. Int. J. Food Sci. Technol. 2003,
38, 285–295. DOI: 10.1046/j.1365-2621.2003.00674.x.
280 Y. B. SAUNSHI ET AL.
[15] Afoakwa, E. O.; Paterson, A.; Fowler, M.; Ryan, A. Flavor
Formation and Character in Cocoa and Chocolate: A Critical
Review. Crit. Rev. Food Sci. Nutr. 2008, 48, 840–857. DOI: 10.
1080/10408390701719272.
[16] Schwan, R. F.; Wheals, A. E. The Microbiology of Cocoa
Fermentation and Its Role in Chocolate Quality. Crit. Rev. Food
Sci. Nutr. 2004, 44, 205–221. DOI: 10.1080/10408690490464104.
[17] Camu, N.; De Winter, T.; Verbrugghe, K.; Cleenwerck, I.;
Vandamme, P.; Takrama, J. S.; Vancanneyt, M.; De Vuyst, L.
Dynamics and Biodiversity of Populations of Lactic Acid
Bacteria and Acetic Acid Bacteria Involved in Spontaneous
Heap Fermentation of Cocoa Beans in Ghana. Appl. Environ.
Microbiol. 2007, 73, 1809–1824. DOI: 10.1128/AEM.02189-06.
[18] Lefeber, T.; Gobert, W.; Vrancken, G.; Camu, N.; De Vuyst, L.
Dynamics and Species Diversity of Communities of Lactic Acid
Bacteria and Acetic Acid Bacteria during Spontaneous Cocoa
Bean Fermentation in Vessels. Food Microbiol. 2011, 28,
457–464. DOI: 10.1016/j.fm.2010.10.010.
[19] Corsetti, A.; Perpetuini, G.; Schirone, M.; Tofalo, R.; Suzzi, G.
Application of Starter Cultures to Table Olive Fermentation:
An Overview on the Experimental Studies. Front. Microbiol.
2012, 3, 248. DOI: 10.3389/fmicb.2012.00248.
[20] de Melo Pereira, G. V.; Miguel, M. G. D. C. P.; Ramos, C. L.;
Schwan, R. F. Microbiological and Physicochemical Characterization
of Small-Scale Cocoa Fermentations and Screening of Yeast and
Bacterial Strains to Develop a Defined Starter Culture. Appl. Environ.
Microbiol. 2012, 78, 5395–5405. DOI: 10.1128/AEM.01144-12.