Skip to main content
NCBI home page
Heliyon logoLink to Heliyon
. 2020 Jan 20;6(1):e03269. doi: 10.1016/j.heliyon.2020.e03269

Microbial synergy between Pichia kudriazevii YS201 and Bacillus subtilis BS38 improves pulp degradation and aroma production in cocoa pulp simulation medium

Honoré G Ouattara a,b,∗, Ryan J Elias b, Edward G Dudley b
PMCID: PMC6971349  PMID: 31993527

Abstract

Interactions between two major microorganisms from Ivorian cocoa fermentation, namely Bacillus subtilis BS38 and Pichia kudriazevii YS201, were investigated during fermentation in cocoa pulp simulation medium. The strains were mutually inhibitory, with Bacillus being more susceptible to this antagonistic effect than Pichia. However, both strains yielded different pulp-degrading enzymes, namely polygalacturonase (PG) from Pichia and pectate lyase (Pel) from Bacillus, that cooperate to efficiently breakdown pectin and vegetable pulp. The quantification of aromas from microbial cultures using Gas Chromatography-Mass Spectroscopy (GC-MS) coupled with headspace microextration (SPME) method, showed that P. kudriazevii produce mainly alcohols such as ethanol (63.165 g/L), phenylethanol (1.005 g/L), methylbutanol (0.138 g/L) and esters, notably ethyl acetate (0.037 g/L) and isoamyl acetate (0.032 g/L). The volatile fraction produced by Bacillus was dominated by butanediol (5.707 g/L), acetoin (1.933 g/L), phenylethanol (0.035 g/L) and acetic acid (0.034 g/L). In co-culture, Bacillus produced low levels of aroma compounds whereas a moderate decrease in the production of these compounds was observed in the yeasts strain. Thus, the dominant aromas present in the co-culture were mainly those from the yeasts strain; however, a 1.37 fold increase of ethanol production was observed in co-culture indicating a synergy between the strains. This study showed that cooperation between B. subtilis BS38 and P. kudriazevii YS201 leads principally to increasing pulp degradation and ethanol production, known as desirable properties for a well processing of cocoa fermentation.

Keywords: Microbiology, Bacillus subtilis, Cocoa fermentation, Ethanol, Interactions, Pectinolytic enzymes, Pichia kudriazevii, Starter culture, Volatile compounds

Microbiology; Bacillus subtilis; Cocoa fermentation; Ethanol; Interactions; Pectinolytic enzymes; Pichia kudriazevii; Starter culture; Volatile compounds.

1. Introduction

Microorganisms play a key role in the fermentation process of cocoa and are known to strongly impact the quality of finished chocolate (Meersman et al., 2016; Ouattara et al., 2008; Schwan and Wheals, 2004). These microorganisms generally belong to the following four groups: yeasts, acetic acid bacteria (AAB), lactic acid bacteria (LAB) and Bacillus. These microorganisms constitute the autochthonous microbiota responsible for natural cocoa fermentation (Ardhana and Fleet, 2003; Lefeber et al., 2010; Ouattara et al., 2011, 2017a,b; Pereira et al., 2012). During fermentation, sugars contained in the pulp are fermented into ethanol and lactic acid by yeasts and LAB, respectively, and AAB further transform ethanol into acetic acid (De Vuyst and Weckx, 2016; Ouattara et al., 2017a,b; Schwan, 1998; Soumahoro et al., 2015). Yeasts, together with Bacillus, produce enzymes that lead to pulp degradation (Cempaka et al., 2014; Ouattara et al., 2008; Schwan et al., 1995; Schwan and Wheals, 2004). This process is essential for the penetration of compounds into beans, principally ethanol and acetic acid. These compounds in turn activate pH-dependent proteolysis enzymes, leading to the degradation of the storage proteins into amino acids and peptides that are aroma precursors of chocolate (Schwan and Wheals, 2004; Voigt and Biehl, 1995; Voigt et al., 1994). A well-developed fermentation leads to a characteristic chocolate flavor and brown color of beans, as well as reducing their bitterness and astringency (Afoakwa et al., 2008, 2015; Aprotosoaie et al., 2016; Beckett, 2009).

Microorganisms are required for proper processing of cocoa into chocolate (Aprotosoaie et al., 2016; De Vuyst and Weckx, 2016; Schwan and Wheals, 2004), and the quality of the final product depends upon the metabolic traits of organisms selected for during fermentation (Crafack et al., 2013; De Vuyst and Weckx, 2016; Schwan and Wheals, 2004). In many parts of the world, the reliance on natural fermentation leads to a process that is difficult to control, resulting in variable quality of cocoa (Lagunes Gálvez et al., 2007). The use of defined starter strains is expected to provide better control over the fermentation process (De Vuyst and Weckx, 2016; Leal et al., 2008; Schwan and Wheals, 2004); however, to date there has been limited success in using defined starter strains to create a product that is superior to those produced by spontaneous fermentations (Lefeber et al., 2012; Crafack et al., 2013; Batista et al., 2016). Previous work often relied on the use of starter strains in pure culture, however here we propose that the limited success with defined starters in cocoa fermentation is due, in part, to a lack of understanding of how these cultures interact during cocoa fermentation. In support of this hypothesis, previous studies have reported that strain performance differs in single versus mixed culture (Qu et al., 2012; Rikhvanov et al., 1999; Van Rijswijck et al., 2017).

The objective of this paper was to analyze the nature of interactions between Pichia kudriazevii YS201 and Bacillus subtilis BS38 which were previously isolated from fermenting cocoa (Ouattara et al., 2011; Samagaci et al., 2016). We hypothesized that the growth rate, production of flavor precursors, and rate of pulp degradation would differ when these organisms were grown together, compared to as single cultures.

2. Material and methods

2.1. Microbial strains and culture medium

The yeast strain P. kudriazevii Ys201 and the bacterium B. subtilis BS38 were previously isolated from fermenting cocoa in Côte d’Ivoire (Ouattara et al., 2011; Samagaci et al., 2016). These strains were stored at -80 °C in 20 % glycerol.

All the microbial cultures were performed using cocoa pulp simulation medium (PSM) (Lefeber et al., 2010; Pettipher, 1986) with slight modifications. The PSM contained 3 % fructose; 2 % glucose; 1 % sucrose; 1 % pectin; 1 % citric acid; 0.6 % yeast extracts; 0.6 % soya peptone; 0.2 % potassium sulphate; 0.2 % magnesium sulphate, 0.04 % manganese sulphate and 0.1 % Tween 80, adjusted to pH 6.0. The concentration for each compound (Sigma Aldrich, Pennsylvania, USA) is indicated as w/v.

2.2. Microbial growth analysis

The growth of strains individually or in co-culture was analyzed in 50 mL PSM medium in a 250 Erlenmeyer flask. Strains for these experiments were first prepared by growing B. subtilis BS38 and P. kudriazevii YS201 in 10 mL nutrient broth (Difco, Detroit, USA) or YPD broth (1 % yeast extract, 1% peptone and 2 % dextrose) respectively, at 30 °C with shaking at 150 rpm. After 12 h incubation, the cultures were diluted 1:2 with PSM and allowed to grow for another 12 h. These precultures were used to seed 50 mL PSM medium at 1:100 (v/v) followed by incubation at 30 °C for 4 days with agitation at 150 rpm. At 3 h intervals, microbial growth was evaluated by reading the absorbance OD600 of the culture using spectrophotometer (Beckman Coulter, model DU-730). For accurate reading of absorbance, cultures were diluted when necessary to have an OD600 below 0.4, as the relationship between OD600 and cell density is linear below this value (Widdel, 2010). Additionally, microbial enumeration was performed by plate count at each growth phase, using the decimal dilution method (Samagaci et al., 2016). The viable cells count (expressed as colony forming units per milliliter; CFU/mL) of P. kudriazevii YS201 was determined after plating on YGC (Yeast Glucose Chloramphenicol), which inhibits B. subtilis, while nutrient agar (NA) supplemented to 100 μg/mL Nystatin (Fisher, USA) was used to selectively quantify B. subtilis.

Microbial growth rate was calculated during exponential phase; it was expressed as:

μ (h−1) = [log (N2) - log (N1)]/ (T2 – T1) where N1 is the number of CFU at time T1 and N2 the number of CFU at time T2 (Dalgaard and Koutsoumanis, 2001). Cultures were duplicate in an experimental set up composed of 6 vials (two vials for each culture). Lag time was estimated by extrapolating the tangent at the exponential part of the growth curve, back to the inoculum level (Dalgaard and Koutsoumanis, 2001).

2.3. Protein purification

For enzymes purification, B. subtilis BS38 and P. kudriazevii YS201 strains were grown in 500 mL of PSM at 30 °C for 48 h. After centrifugation (5,000 × g) for 10 min, the supernatant was subjected to ammonium sulphate precipitation at 100 % saturation. The precipitated proteins were next dissolved in 10 mL distilled water and dialyzed overnight at 4–8 °C in a 14 kDa molecular weight cut off (MWCO) dialysis bag (Fisher, USA). The dialysate from B. subtilis BS38 supernatant was applied to fast flow CM Sepharose cation exchange resin (GE Healthcare, USA) previously equilibrated with 0.1 M Tris-HCl buffer, pH 7.2, whereas the dialysate from P. kudriazevii YS201 supernatant was applied to fast flow DEAE, Sepharose anion exchange resin (GE Healthcare, USA), equilibrated with 0.1 M citrate phosphate buffer pH 6.8. Chromatography was performed on 25 mL bed volume of resin contained in a gravity Econo column®, 30 cm length, 2.5 cm internal diameter (Bio-Rad, USA). The different protein fractions (10 mL) were eluted with increasing concentrations of NaCl from 0 to 500 mM. The purity of enzymatic proteins was assessed by electrophoresis using SDS and 12 % polyacrylamide gel. Molecular markers were used to approximately calculate the molecular weight of the proteins (Bio Rad Precision Plus Protein™, USA). Protein bands were detected by Coomassie blue staining, and the concentration was estimated using the Bradford reagent (Sigma Millipore, Bellefonte, USA).

2.4. Enzymatic assays

Pectinolytic enzymes were assayed in 1 mL of reaction mixture containing 0.2 % (w/v) apple pectin (Sigma Aldrich, Pennsylvania, USA), and 0.1M Tris-HCl pH 8 or citrate-phosphate pH 5. The reactions were run for 1 h, at 40 °C and enzymatic activities were measured spectrophometrically (OD540) following the release of reducing group from pectin degradation using the dinitrosalicylic (DNS) acid method (Miller, 1959). One unit of polygalacturonase or pectate lyase was defined as the amount of enzyme that releases 1 μmol of galacturonic acid (GA) per minute, under experimental conditions.

2.5. Thin layer chromatography (TLC)

TLC was performed as described previously by Lojkowska et al. (1995). Pectin was digested by each purified enzyme at 0.05 μg/ mL, in a 100 μL reaction mixture as described above. Incubation was performed overnight at 45 °C. Next, 20 μL of the reaction products were applied on a silica gel chromatography sheet (10 cm Width, and 20 cm height) (Merk, Germany) and separated for 5 h using a mobile phase composed of n-butanol – acetic acid – water (5: 2: 3). The sheet was air dried at room temperature, and then immerged in a revelation solution containing phosphomolybdic acid - sulfuric acid – ethanol, (3: 10: 87). Finally, the sheet was plate heated at 100 °C to reveal the spot of the reaction products from pectin degradation.

2.6. Pulp degradation and maceration capacity of enzymes

The ability of microbial pectinolytic enzymes to macerate vegetable was analyzed using the epicarp and mesocarp of cucumber (Cucumis sativus) as solid substrate. Cucumber is a standard model generally used for studying native pulp degradation and maceration capacity of pectinolytic enzymes from microbial culture (Schwan et al., 1997; Khan and Latif, 2016). Each sample (of cucumber) weighting approximately 5 g was submerged in 10 mL of crude enzyme from the 48 h culture supernatant described above. The submerged sample was incubated for 24 h at 30 °C under shaking at 150 rpm. The negative control consisted of cucumber submerged in PSM (enzyme free). After incubation, the crude enzyme or the PSM was removed, replaced with 10 mL distilled water, and then vortexed at highest speed (VWR, Analog Vortex Mixer) for 1 min. The debris obtained from the macerated cucumber was centrifuged at 100 × g for 5 min.

Pectin degradation was also measured after incorporation of 0.2 % (w/v) apple pectin into 20 % agar, at pH 5.0 (citrate-phosphate buffer) or pH 8.0 (Tris-HCl, buffer). After solidification of the agar, 5 U of enzyme was spotted on pectinized plate agar, incubated 5 h at 30 °C. Then, the solid plate was flooded with lugol solution (Sigma Aldrich, Pennsylvania) to detect clearance zones corresponding to enzymatic activity (Ouattara et al., 2008).

2.7. Volatiles analysis

Volatile compounds produced by the strains were analyzed under single and co-culture conditions. B. subtilis BS38 and P. kudriazevii YS201 were grown at 30 °C for 18–24 h on NA and YPD agar, respectively. Colonies were picked and resuspended in sterile distilled water to an OD600 = 0.05. This suspension was used to inoculate to 1% (v/v), 5 mL of PSM medium contained within a 20 mL sterilized vial suitable for GC-MS analysis (Restek, USA). The vial was sealed with an 18 mm cap with a thermo-resistant silicon septum, and was incubated at 30 °C for 48 h with constant agitation at 150 rpm in orbital shaking incubator. The negative control was non-inoculated medium incubated under the same conditions. Cultures were prepared in triplicate.

The extraction of the volatile compounds was performed using solid phase microextraction of the vial headspace (SPME-HS) using a 50/30 μm divinylbenzene/ carboxene/ polydimethylsiloxane (DVB/CAR/PDMS) fiber (Supelco,/Sigma Aldrich, Bellefonte, PA, USA) as previously described by Koné et al. (2015). The fiber was previously conditioned in the chromatograph injector at 250 °C for 3 min and then was exposed to the sample headspace at 50 °C for 45 min. Extracted volatile compounds were analyzed using an Agilent Technologies 7890A gas chromatograph (Agilent Technologies, Inc., Santa Clara, CA) equipped with a Gerstel MPS2 multipurpose autosampler (Gerstel GmbH & Co. KG. Mülheim an der Ruhr, Germany) that was coupled to an Agilent 5975C mass selective detector that was a Hewlett Packard DBWAX capillary column (30 m length × 0.25 mm internal diameter × 0.25 μm film thickness) (Agilent, Palo Alto, CA, USA). Agilent MassHunter GC/MS Acquisition software (Version B.07.00 SP2.1654) and Gerstel Maestro 1 software (Version 1.3.20.41) were used to control the instrument and autosampler, respectively.

The GC oven temperature was initially set at 40 °C, was held for 5 min, and then increased to 140 °C at a rate of 2 °C/min. Helium was used as the carrier gas at a flow rate of 1 mL min−1. Headspace samples were injected in splitless mode over the course of 2 min and the injector temperature was held at 250 °C.

Volatile compounds were putatively identified by probability based matching of their mass spectra with those obtained from a commercial database (Wiley275.L, HP product) and by matching the Kovac index (KI) of the compounds with literature values reported for equivalent columns. Whenever possible, the identification was confirmed by using pure reference standards of the components.

Relative quantification was performed from peak areas integration of well-identified compounds using the MSD Chemstation software (version E.02.02.1431, Agilent Technologies). Absolute quantification was performed using external and internal standards. External standard curves were obtained with good linearity over the relevant concentration range of the sample volatiles (R2 0.997–0.999). Internal standards consisting of isotopically labeled forms of the compounds were used at constant concentrations to normalize the differences in the response of the GC detector in the vials (see supplemental material Table S1). Volatile concentrations were calculated based on linear regression equations derived from their standard curves.

2.8. Statistical analysis

When required, a one-way analysis of variance (ANOVA) using StatPlus software version 5.4 at 95 % confidence was performed to determine whether there are any significant differences between the means of independent group of replicates.

3. Results

3.1. Co-culture negatively impacts the growth of both B. subtilis and P. kudriazevii

To analyze the interaction between B. subtilis and P. kudriazevii during co-culture, growth was compared between single- and co-culture. The initial OD600 for the single cultures was 0.017 for B. subtilis BS38 and 0.065 for P. kudriazevii YS201 and the OD600 from co-culture was 0.089, corresponding approximately to the sum of absorbances from the single cultures (Table 1). During the lag phase, the sum of OD600 from single cultures corresponded to the OD600 obtained in co-culture whereas differences increased between these two values during the rest of growth cycle suggesting decreased microbial growth in co-culture (Figure 1).

Table 1.

Microbial growth during fermentation in the different cultures.

Growth phases
 Single culture
B. subtilis
 Single culture
P. kudriazevii
 Coculture
 Theoretical expected growth in co-culture
OD600 Log (CFU/mL) OD600 Log (CFU/mL) OD600 B. subtilis Log (CFU/mL) P. kudriazevii Log (CFU/mL) OD600
Lag Start 0.017 (±0.004) 5.02 (±0.045) 0.065 (±0.011) 5.38 (±0.15) 0.089 (±0.006) 5.02 (±0.045) 5.38 (±0.15) 0.082 (±0.015)
End 0.05 (±0.008) 5.62 (±0.08) 0.158 (±0.019) 5.84 (±0.115) 0.212 (±0.019) 5.60 (±0.071) 5.79 (±0.095) 0.208 (±0.027)
Exponential End 1.85 (±0.122) 7.24 (±0.075) 8.16 (±0.225) 7.60 (±0.055) 8.64 (±0.201) 7.10 (±0.105) 7.52 (±0.091) 10.01 (±0.347)
Stationary Late 3.16 (±0.138) 7.47 (±0.065) 10.16 (±0.45) 7.69 (±0.085) 10.85 (±0.960) 7.21 (±0.053) 7.59 (±0.085) 13.32 (±0.588)
Growth rate μmax(h−1) 0.373a (±0.025) 0.405b (±0.013) 0.345c (±0.008) 0.398d (±0.011)
Open in a new tab

Duration (start lag phase – end lag phase) = 2 h; (end lag phase – end exponential phase) = 10 h; (end exponential phase – late stationary phase) = 10 h; μmax maximum growth rate measured during exponential phase. In the same row, numbers with different letters are significantly different. Cultures were duplicate in an experimental set up composed of 6 vials (two vials for each culture).

Figure 1.

Figure 1

Open in a new tab

Comparative evolution of microbial growth in singles and co-culture. Red line is the sum of OD600 from the different single cultures as theoretical growth expected in co-culture. Blue line is the actual and experimental microbial growth measured in co-culture. Arrows indicate the time at which viable cells enumeration on plate agar was performed (see Table 1). (a) lag phase, (b) exponential phase, (c) stationary phase.

We further quantified the growth of each culture by performing viable cell counts on selective agar. We observed that the growth of B. subtilis BS38 in co-culture (μmax of 0.345 ± 0.008 h−1) was slower than growth in single culture (μmax of 0.373 ± 0.025 h−1) (Table 1). The viable cell counts of B. subtilis BS38 at the end of the exponential growth phase in co-culture was 7.1 ± 0.105 log CFU/mL thus representing 72 % of the bacterial load obtained in single culture, which was 7.2 ± 0.075 log CFU/mL. Likewise, the growth rate of P. kudriazevii YS201 decreased from 0.405 (±0.013) h−1 in single culture to 0.398 (±0.011) h−1 in co-culture (Table 1). The growth of this strain in co-culture reached 7.52 ± 0.091 log CFU/mL thus representing 83 % of the load obtained in single culture (7.6 ± 0.055 log CFU/mL). For both strains, the growth rates were significantly different in the two conditions (single and co-culture) as assessed by ANOVA.

3.2. Microbial growth analysis at different pH

Analysis of microbial growth in acidic range showed that at pH 4, B. subtilis reach only 4.82 % of the maximum OD600 obtained at pH 6 (Figure 2A). This poor growth of Bacillus was characterized by a long lag phase of 10 h and maximum OD600 of 0.168 (±0.021) (Figure S1). When, these Bacillus cells were transferred and grown at pH 6, they recovered their growth capacity. On the other hand, P. kudriazevii YS201 at pH 4, grew to a final OD600 that was approximately 60 % of that of cells grown at pH 6 (Figure 2A).

Figure 2.

Figure 2

Open in a new tab

Growth of strains analyzed at different pH. A. Effect of pH on the growth of strains. OD600 obtained at the end of the cultures after 24 h, 30 °C, in PSM are shown. The relative growth was expressed as the percentage of absorbance using maximum OD600 at pH 6 as 100 %. B. Level of P. kudriazevii growth reached in single and co-culture at different pH. The relative growth was calculated as the percentage of viable cells (VC) obtained in co-culture using the load of VC in single culture as 100 %.

Given these results, two co-culture conditions were analyzed: co-culture at pH 4 where Bacillus and Pichia growth differed dramatically, and co-culture at pH 6 where both strains grew well. At pH 4, P. kudriazevii co-cultured with B. subtilis reached 7.42 ± 0.105 log CFU/ mL after 24 h, thus representing 94.79 (±3.33) % of the density observed in single culture (Figure 2B), while the load of Bacillus in this co-culture was 6.07 ± 0.055 log CFU/ mL.

At pH 6, the growth of P. kudriazevii in the co-culture achieved a load of 7.55 ± 0.085 log CFU/ mL thus representing 78.15 (±4.78) % (Figure 2B) of cell concentration reached in single culture. Bacillus reached 7.17 ± 0.06 log CFU/ mL in this co-culture.

On the other hand, whatever the initial pH of the medium, it did not significantly vary at the end of the cultures. The more notable shift of pH was 6.07 (from initial pH 6.0), observed in Bacillus single culture and co-cultures.

3.3. Basic characteristic of purified pectinolytic enzymes

Enzymes from both strains were first purified to electrophoretic homogeneity. During the purification process, pectin degrading activity from B. subtilis BS38 weakly bound on CM sepharose, and eluted with a weak ionic strength of salt solution 0.1 M NaCl while a higher concentration of NaCl (0.45 M) was required to elute the enzyme produced by P. kudriazevii YS201 from DEAE sepharose. The enzyme from B. subtilis BS38 has a molecular weight of approximately 25 kDa, and split pectin to give products that strongly absorbed in the UV range (230 nm) (Figure S2). Addition of 0.1 mM calcium to the reaction mixture strongly increases the activity of this enzyme showing calcium-dependence. Moreover, this enzyme was completely inhibited by 0.1 mM EDTA (Table S2). The enzyme from B. subtilis was therefore identified as a pectate lyase (Pel) (EC 3.2.1.15) that showed maximum activity at pH 8 in liquid phase reaction, with a large range of activity (pH 6.5–9) but total loss of activity below pH 6 (Figure S3). In contrast, pectinolytic enzyme from P. kudriazevii YS201 was identified as polygalacturonase (PG) (EC 4.2.2.2). This enzyme yields products that did not absorb light in the UV range or be inhibited by EDTA (Table S2). The purified PG had a molecular weight of approximately 31 Kda, a pH range of activity between 3 and 5.5 (Figure S3). No other pectinolytic enzyme such as pectin methyl esterase was found in the supernatants.

Results from thin layer chromatography showed that, PG and Pel individually at their respective optimum pH (5 and 8) degrade pectin; monomer was not detectable after digestion regarding the spot a, that was not present in the lane 2 and 3 (Figure 3). At pH 8, it was observed that Pel led to a sharp degradation of pectin represented by the spot f, which intensity was visibly weak (Figure 3, lane 3). In comparison, more pectin remained after PG digestion (pH 5) regarding the visible intensity of spot f (Figure 3, lane 2). Moreover, the breakdown of the pectin led to accumulation of degradation products that were more important with Pel action as showed the visible intensity of the spots b, c, d and e, lane 3, comparatively to PG action, lane 2 (Figure 3). When PG and Pel were mixed and incubated with pectic substrate in the same buffer at pH 6, no pectin degradation products were observed as the spot b, c, d, and e were not present in lane 4 (Figure 3). Indeed, the combined enzymes failed to degrade pectin at pH 6 in liquid phase.

Figure 3.

Figure 3

Open in a new tab

Thin layer chromatography showing the degradation products from pectin due to enzymes activity. a, b, c, d and e are different products from pectin degradation; f is the polymeric pectin that remained unmoved from the initial line during chromatography. Arrow indicates the migration direction. Pectin was incubated with 1.5 U of enzyme and incubated at 45 °C overnight. Lane 1 monomeric galacturonic acid used as marker; lane 2, PG in 0.1 M acetate reaction buffer pH 5; lane 3 Pel in 0.1 M Tris-HCl reaction buffer pH 8; lane 4, mixed PG and PL in 0.1 M acetate reaction buffer pH 6. The non-degraded pectin (spot f) (lane 4) remains abundant comparatively to the pectin (lane 3) effectively degraded.

3.4. Cooperative action of P. kudriazevii YS201 and B. subtilis BS38 leads to strong degradation of pectin

Due to their different characteristics at different pH (Figure S2), it was not possible to find a common active pH at which the combined action of these enzymes could be analyzed in liquid phase. We next tested activity on a solid support consisting of a buffered solution of pectin solidified with agar. The results show that Pel from B. subtilis BS38 was capable of degrading pectin on solid phase even at pH 5 (data not shown). Similarly, PG from P. kudriazevii YS201 was still active at pH 7.5 on solid phase. Therefore, the combined action of these enzymes was analyzed on pectin concentrated solid phase (2 %) at pH 6. The results showed that a mixture of both enzymes (Pel + PG) lead to a sharp degradation of pectin as showed the visible clear halos, comparatively to individual enzymes (Figure 4).

Figure 4.

Figure 4

Open in a new tab

Cooperative degradation of pectin by different purified enzymes on solid phase. 5 U of enzymes were dropped on the dried solid phase containing 2 % pectin, 0.1 M acetate buffer pH 6, and incubation for 5 h at 30 °C. A: 2.5 U of each enzyme solution were mixed (Pel from B. subtilis + PG from P. kudriazevii) B: 5 U of Pel C: 5 U of PG.

We further analyzed this cooperative effect by testing the supernatant directly from co-culture, using cucumber as a substrate. A complete disintegration of the vegetal tissue was achieved after 48 h contact with the co-culture supernatant whereas only a softening was obtained with the single culture supernatants (Figure 5).

Figure 5.

Figure 5

Open in a new tab

Maceration capacity of different crudes enzymes. A: crude enzyme from B. subtlis, B: crude enzyme from P. kudriazevii D: crude enzyme from co-culture diluted ½ with distilled water. C: control.

3.5. General profile of aroma compounds yielded by B. subtilis BS38 and P. kudriazevii YS201

Analysis of volatiles produced by the strains in different cultures was carried out in two steps. In the first step, the compounds were identified based on the retention time and in the next step, we performed an absolute quantification of the analyts using standard curves. To avoid misinterpretation, volatile compounds detected in non-inoculated medium (negative control) were discarded from data analysis (Table 2).

Table 2.

Endogenous volatile compounds present in the medium.

Volatile in the PSM medium
2-methyl propanal butanal
2-methyl butanal
3-methyl butanal
n-hexanal
furaldehyde
nonanal
benzaldehyde
Open in a new tab

Non-inoculated medium contained in 20 mL sterilized vial (Restek, USA) tightly closed was incubated at 30 °C for 48 h, under shaking at 150 rpm. Volatiles were detected using GC-MS and identified by probability based matching of their mass spectra with those obtained from a commercial database (Wiley275.L, HP product) and by matching the Kovac index (KI) of the compounds with literature values reported for equivalent columns.

The results indicated that the observed volatile compounds generated by P. kudriazevii YS201 were mainly alcohols and esters while acids were not detected (Table 3). Alcohols were mainly ethanol, methyl butanol and phenylethanol while ethyl acetate, isoamyl acetate and phenyl acetate composed the class of ester produced by this yeast strain (Table 3).

Table 3.

Volatile compounds produced by P. kudriazevii strain studied.

Volatile compound detected Relative abundance (%) Odor description References
Alcohol Ethanol 23.055 (±2.547) – –
1-Butanol, 3 methyl 13.284 (±3.893) Bitter, chocolate Rodriguez-Campos et al., 2012
1-Butanol, 2 methyl 6.068 (±0.956) Fruity Ramos et al., 2014
Phenyl alcohol 28.961 (±8.212) Floral Rodriguez-Campos et al., 2012
Acids Isovaleric acid* – Sweat, rancid Rodriguez-Campos et al., 2012
Acetic acid* – Sour, vinegar Rodriguez-Campos et al., 2012
2 Methyl butanoic acid* – Sweaty Krings et al., 2006
Esters Methyl acetate – – –
Ethyl acetate 22.048 (±2.529) Fruity pineapple Rodriguez-Campos et al., 2012
Ethyl propanoate 0.962 (±0.422) –
Isoamyl acetate 4.383 (±1.505) Fruity, banana Ramos et al., 2014
Phenyl ethyl acetate 1.235 (±0.285) Fruity sweet Rodriguez-Campos et al., 2012
Isobutyl acetate* – Fruity Rodriguez-Campos et al., 2012
Ethyl butanoate* – – –
Others Methionol 3 methyl thiol propanol* – – –
Open in a new tab

Culture contained in 20 mL sterilized vial (Restek, USA) tightly closed was incubated at 30 °C for 48 h, under shaking at 150 rpm. For GC-MS analysis, the relative abundance of volatile compounds in each culture was calculated using area integration from chromatogram. Areas were normalized by using total areas from all volatiles compounds detected in the culture as 100 %. Standard deviations were obtained from three replications. The compounds constantly detected in all the cultures were considered for abundance calculation. (*) Compounds not constantly detected in the cultures. ± Standard deviation between three biological replications.

With respect to B. subtilis BS38, this strain produced a wider spectrum of volatile compounds that belong to the class of alcohols, ketones, acids, pyrazines and furans (Table 4). Alcohols presented a large proportion of these volatiles compound from B. subtilis BS38 with the yield of phenyl ethanol, 1-butanol, 3-methyl butanol, 2-methyl butanol and 2,3-butanediol. On the other hand, 2,3-butanedione and 3-hydroxy 2-butanone represented the class of ketones produced whereas acetic acid and dimethyl ethyl pyrazine composed the other class of volatile compound produced from this Bacillus strain (Table 4).

Table 4.

Volatile compounds produced by B. subtilis strain studied.

Volatile compounds detected Relative abundance (%) Odors description References
Alcohols 1-Butanol 1.401 (±0.075) – –
1-Butanol, 3 methyl 6.008 (±0.823) Malty, chocolate Rodriguez-Campos et al., 2012
1-Butanol, 2 methyl 2.579 (±0.489) Fruity Ramos et al., 2014
2, 3-Butanediol 47.458 (±10.260) Sweet, chocolate Ramos et al., 2014
Ethyl hexanol* – Fruity, green Bonvehí, 2005
Phenol* – Smoky Rodriguez-Campos et al., 2012
Phenyl alcohol 1.755 (±0.476) Floral Rodriguez-Campos et al., 2012
Ketones 2,3 Butadione 5.770 (±1.019) Buttery Rodriguez-Campos et al., 2012
3-Hydroxy, 2 butanone 28.432 (±7.572) Buttery, creamy Rodriguez-Campos et al., 2012
Acids Acetic acid 5.267 (±1.837) Sour, vinegar Rodriguez-Campos et al., 2012
Esters Isoamyl acetate* – Fruity, banana Ramos et al., 2014
Pyrazines Dimethyl ethyl pyrazine 1.325 (±0.115)
Others Methionol 3 methyl thiol propanol – – –
Furan methanol – Cooked sugar Bonvehí, 2005
Furan, 2 phenyl – – –
Furan, 2,5 dimethyl – – –
Propan, 2 chloro, 2 nitro – – –
Dimethyl disulfide – Sulfurous Bonvehí, 2005
Open in a new tab

Culture contained in 20 mL sterilized vial (Restek, USA) tightly closed was incubated at 30 °C for 48 h, under shaking at 150 rpm. For GC-MS analysis, the relative abundance of volatile compounds in each culture was calculated using area integration from chromatogram. Areas were normalized by using total areas from all volatiles compounds detected in the culture as 100 %. Standard deviations were obtained from three replications. The compounds constantly detected in all the cultures were considered for abundance calculation. (*) Compounds not constantly detected in the cultures. ± Standard deviation between three biological replications.

Some compounds were produced only by one strain. For instance, ethanol and all esters produced by P. kudriazevii YS201 were not observed in B. subtilis BS38 culture medium. Likewise, compounds such as butanol, butanediol, 2, 3-butanedione as well as 3-hydroxy 2 butadione and dimethyl ethyl pyrazine produced by B. subtilis BS38 were not be detected in P. kudriazevii YS201 cultures. On the other hand, the results showed that some compounds (e.g., phenyl alcohol, methylated butanol) were produced by both microbial strains. In co-culture the abundance of volatile compounds was different from that observed in single culture.

3.6. Interactions of strains in co-culture lead to the modification of aroma compounds production

Quantification analysis of the volatiles revealed that the studied microbial strains yielded an array of aroma compounds at concentrations above the known odor threshold (Table 5). In single culture, P. kudriazevii YS201 was able to produce up to 63.165 (±6.536) g/L ethanol. This production increased significantly in co-culture where it reached 86.596 (±4.177) g/L (Table 5). However, for the other volatiles, a general decrease was observed in co-culture. For instance, the production of phenylethanol and 3-methylbutanol, two of the main volatiles present in P. kudriazevii YS201 single culture, were observed at concentrations of 1.005 (±0.082) and 0.138 (±0.042) g/L, respectively. These levels in co-culture decreased to 0.482 (±0.022) and 0.116 (±0.017) g/L, respectively (Table 5 and Figure 6). Moreover, the decrease of aroma compounds production comparing single to co-culture was most dramatic with B. subtilis BS38, which produced butanediol (5.707 ± 0.098 g/L), 3-hydroxy 2-butanone (acetoin) (1.933 ± 0.221 g/L) and acetic acid (0.034 ± 0.004 g/L) in single culture but almost failed to produce these compounds in co-culture.

Table 5.

Absolute quantification of the main aroma produced by strains in single and co-culture.

Compounds Aroma production (g/L)
 Odor threshold (g/L)
B. subtilis P. kudriazevii B. subtilis + P. kudriazevii
Ethyl acetate nd 0.037 (±0.002) 0.042 (±0.008) 0.006 National Institute of Health, 2015
Isoamyl acetate 0.004 (±0.000)a 0.032 (±0.008)b 0.020 (±0.003)c 34 × 10−5National Institute of Health, 2015
Ethanol nd 63.165 (±6.536) 86.596 (±4.177) 0.049 National Institute of Health, 2015
Butanediol 5.707 (±0.098) nd 1.102 (±0.200) –
Phenyl ethanol 0.035 (±0.003)a 1.005 (±0.082)b 0.482 (±0.022)c 0.007 National Institute of Health, 2015
2 methyl butanol nd 0.042 (±0.001) 0.032 (±0.004) 12 × 10−5National Institute of Health, 2015
3 methyl butanol 0.005 (± 0.000)a 0.138 (±0.042)b 0.116 (±0.017)c 0.3 × 10−5Bonvehí, 2005
3 hydroxy 2 butanone 1.933 (±0.221) nd nd –
Acetic acid 0.034 (±0.004) nd nd 0.032 Bonvehí, 2005
Open in a new tab

nd, not detected. Absolute quantification was performed using external and internal standards. Different concentrations of external standards allowed plotting standard curves. Internal standards were composed of isotope labeled form of the compounds that were used at constant concentrations to normalize the differences in the response of the GC detector in the vials. Then the concentrations were calculated from the linear regression equation of the standard curves. ± Standards deviation between three biological replications. In the same row, numbers with different letters are significantly different. NIH, National Institute of Health, (USA).

Figure 6.

Figure 6

Open in a new tab

Comparative levels of aroma production in single and co-culture. Relative production was expressed as percentage, using the quantity of aroma produced in single culture as 100 %. Error bars indicate standard deviation between three replicates.

It was also observed that the quantity of volatile compounds (e.g., isoamyl acetate, phenylethanol, methylbutanol) produced by both strains in single culture did not increase in the co-culture treatment (Table 5). Figure 6 shows the level of volatiles production in single and co-culture. Similar modifications in volatiles production in co-culture were also observed at 35 °C, the optimum growth temperature for both strains (data not shown).

Since the pH was found to influence the interaction in co-culture during microbial growth, the productions of the main volatile compounds from P. kudriazevii YS201, notably ethanol and phenylethanol, were analyzed in various cultures differing in the medium pH. Figure 7 shows that, the production of both compounds was similar in single and co-culture at pH 3 and 4 whereas strong modifications in volatiles production in co-culture occur at pH 5 and 6.

Figure 7.

Figure 7

Open in a new tab

Ethanol production by P. kudriazevii in single and co-culture at different pH. A. Ethanol, B phenylethanol. Relative production was expressed as percentage, using the quantity of aroma produced in single culture as 100 %. Error bars indicate standard deviation between three replicates.

4. Discussion

Interaction between two major microbial species notably P. kudriazevii YS201 and B. subtilis BS38, involved in Ivorian cocoa fermentation was investigated. Most of previous studies of microbiota from fermenting cocoa focused on strain identification and typing (Nielsen et al., 2007; Ouattara et al., 2017b), biochemical and functional properties (Pereira et al., 2013), and more recently the use of microbial cultures as starters in experimental cocoa fermentation (Leal et al., 2008; Meersman et al., 2016). Apart from the study describing an antifungal activity of a co-culture of yeast and lactic acid bacteria (Romanens et al., 2019), the direct interactions between different microbial species from fermenting cocoa have rarely been investigated. The present work provides new understanding of how microbial strains work together during fermentation of cocoa.

The pulp simulation medium (PSM) used in this study, although modified, allowed a good growth of the microorganisms analyzed that were able to reach an OD600 up to 10 for P. kudriazevii and 3 for B. subtilis, within 24 h. However, this absorbance corresponded to a relatively low count of viable cells that never reached 8 log (CFU/mL). In comparison to standards, an OD600 of 10 from yeast culture and 3 from Bacillus are expected to give respectively 8.47 and 9.17 log CFU/mL (Day et al., 2004; Fisher et al., 2001). Indeed many factors such as composition and pH of the medium are known to influence the OD of microbial culture.

Moreover, the maximum load of Bacillus and yeast obtained during natural fermentation of cocoa can sometime reach 9 log CFU/mL (Ardhana and Fleet, 2003; Nielsen et al., 2007; Schwan, 1998) that is higher than the growth obtained in PSM. This suggests that the PSM may be less favourable for microbial growth than the natural cocoa pulp. As to support this suggestion, we isolated in previous study B. subtilis strains throughout the fermentation process of cocoa even during the first stage of this process when pH was round 3.5 (Ouattara et al., 2008, 2011). Yet, in the present study, B subtilis BS38 as one of these strains was not able to grow in PSM at pH 3.5, and very poorly at pH 4.

Analysis of microbial growth at different pH in PSM shows that in co-culture, at pH 4, the yeast strain grew similarly to a single culture, and do not appear to be hindered by the presence of Bacillus that grow poorly indicating weak interaction. However, at pH 6 corresponding to optimum growth for both strains, a significant decrease of microbial population was observed for both strains growing in co-culture, indicating an antagonistic type of interactions. Such interactions could be responsible for differences in end products observed. For instance, phenylethanol and ethanol produced in single culture by P. kudriazevii YS201 did not differed significantly to that produced in co-culture at pH 3 and 4 at which B. subtilis BS38 had a poor growth. However, the remarkable variations in the metabolite production observed at pH 5 and 6 in co-culture may reflect strong interaction between strains. These interactions are also expected to occur at temperatures range in which both strains can grow well notably 35–40 °C as shown previous results (Samagaci et al., 2014; Yao et al., 2017). Such interactions leading to significant variations of metabolites production in microbial strains are undoubtedly critical since they could be exploited to increase certain metabolites of interest like ethanol and tune the quality of chocolate. However, these interactions could also present the disadvantage of lowering the level of certain aroma compounds such as phenylethanol as revealed this study.

As PSM may be less favourable for microbial growth than the natural cocoa pulp, results may not directly translate to the fermentation of cocoa. However, this may still provide insights into processes occuring in natural cocoa fermentation where microbial succession is characterized by a growth of yeast population to 9 log CFU/ mL during the 24 first hours, increasing pH from 3.5 to 4, while Bacillus load remains around 7 log CFU/ mL (Ardhana and Fleet, 2003; Pereira et al., 2012). This trend typically reverses in later stages of the fermentation when pH reaches 6, allowing Bacillus to reach up to 8 log CFU/mL while yeast decline to roughly 3 log CFU/mL (Ardhana and Fleet, 2003; Schwan, 1998). In this context, interactions are expected to occur during the period where both strains are present with relatively high load notably the first four days of cocoa fermentation (pH round 3.5–4.5 and temperature 30–35 °C) (Ardhana and Fleet, 2003; Schwan and Wheals, 2004; Kouame et al., 2015). According to Gibe and Pangan (2015) reported that the use of Bacillus and Pichia starter cultures proved to be efficient within this period of first four days. Another shortcoming of the PSM system used here, is the lack of lactic acid bacteria and acetic acid bacteria, which may alter how Bacillus and yeast interact in natural fermentation of cocoa.

Pulp-degrading enzymes are important to cocoa fermentation due to their ability to speed the fermentation process (Bhumibhamon and Jinda, 1997; Cempaka et al., 2014). Thin layer chromatography (TLC) was performed to further analyze the cooperative effect of Pel and PG and to define the contribution of each specific pulp-degrading enzymes. This study found that Pel and PG, at their respective optima pH, individually generated a mixture of different size oligomers from pectin degradation. The absence of monomers among these pectin degradation products indicates that these enzymes lack exopectinase activity, and thus hydrolyze pectin randomly, leading to the same profile of end-products as shown by TLC. Unfortunately, it was not possible to check whether the chromatogram from combined action of Pel and PG is different from that of individual enzyme activity as these enzymes failed to degrade pectin at a same pH in liquid phase.

In contrast, synergistic degradation of pectin by these enzymes was observed on solid phase (pectinized plate agar and cucumber assays). These experiments may better mimic cocoa fermentation conditions, where pulp is the solid substrate from which water has drained at the beginning of the fermentation. This data also argues that enzymes activities measured from commonly used liquid phase may not accurately depict what is occurring during cocoa fermentation. It is generally reported that, surface adhesion-based culture such as solid-state fermentation (SSF) including microbial cells growing on a solid phase (biodegradable or not), produce more pectinase than culture in liquid medium (Submerged fermentation) (Kumar et al., 2011; Oumer and Abate, 2018). Our experiments show this may also be true for pectinase activity, shown here to be more efficient on solid phase (plate agar or cucumber) than in liquid. The cooperation of pulp-degrading enzymes from both strains P. kudriazevii YS201 and B. subtilis BS38 observed in this study, may also play a determinant role in speeding the fermentation process of cocoa. This is strongly supported by recent studies, reporting that a mixture of yeast and Bacillus species namely Pichia mexicana and Bacillus pumilus was successful in shortening fermentation process from 6 to 4 days while single treatment of either P. mexicana or B. pumilus was insufficient to ferment cacao beans within 4 days (Gibe and Pangan, 2015). Cucumber, as a standard model for studying the pulp-degrading capacity of enzymes (Schwan et al., 1997; Khan and Latif, 2016) proved to be suitable in our study since this cooperative effect of Pel and PG could not be revealed using the cocoa pulp itself.

Previously, it was reported that microorganisms could impact the aroma of the final chocolate by directly producing aroma compounds during fermentation of cocoa (Meersman et al., 2016). In the present study, aromas produced by Pichia in PSM belonged essentially to the class of alcohol namely ethanol, phenylethnol and methylbutanol and the class of esters notably ethyl acetate and isoamyl acetate. These compound were also found to be among the main aromas produced by yeast strain from fermenting cocoa, using saboureau medium as reported Koné et al. (2015), but without quantification of aroma compounds produced. Moreover, to the best of our knowledge, apart from a former study reporting the production of pyrazines (Zak et al., 1972), the volatile compounds production of Bacillus from fermenting cocoa, susceptible to modulate aroma quality of final chocolate has been rarely investigated. Here B. subtilis strain BS38 showed the ability to generate an array of aroma compounds, mainly acetoin, butanediol, methylbutanol and pyrazines. However, in co-culture, volatiles production by yeast decreased while Bacillus almost failed to produce its’ potential of aroma compounds. This confirms the interactions between these strains in fermenting medium where volatiles from yeast should dominate the aroma from Bacillus.

The increase of ethanol production observed in co-culture during this study is indicative of a synergistic action of both strains. Likewise, B. subtilis and Saccharomyces cerevisiae were also used in co-culture to improve ethanol yield from wastewater (Tantipaibulvut et al., 2015). Generally, this synergy is thought to result from polysaccharides hydrolysis by Bacillus yielding metabolic substrates that are converted by yeast to ethanol. The synergy between Bacillus and yeast strain for ethanol production may play a crucial role since, this metabolite is highly required for a well processed cocoa fermentation (De Vuyst and Weckx, 2016).

In conclusion, this study evidenced interactions between two microbial species during fermentation of cocoa pulp and suggests the probable existence of other interactions between other microorganisms in the fermenting cocoa. It also opens new perspectives of research on microbial interactions that may help for better understanding of the exact role of the microbiota in fermenting cocoa and for better manipulation of starter strain to improve this fermentative process.

Declarations

Author contribution statement

Honoré G. Ouattara: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Ryan J. Elias: Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Edward G. Dudley: Conceived and designed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Funding statement

This work was supported by a Fulbright grant (African Scholar Research) from the US State Department awarded to Honoré G. Ouattara under the grant number E0581134. Edward Dudley and Ryan Elias are supported in part by USDA National Institute of Food and Agriculture Federal Appropriations under Project PEN04522 and Accession number 0233376.

Competing interest statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.

Acknowledgements

The authors kindly acknowledge Dr. Wai Fun Leong for her technical assistance. The authors gratefully acknowledge Dr. Olga Shimelis from Millipore Sigma, Bellefonte, PA (USA) for providing the standards used in this study for aromasidentification and quantification.

Appendix A. Supplementary data

The following is the supplementary data related to this article:

Supplementary materials
mmc1.docx (96.2KB, docx)

References

  1. 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. [DOI] [PubMed] [Google Scholar]
  2. Afoakwa E.O., Kongor J.E., Budu A.S., Mensah-brown H., Takrama J.F. Changes in biochemical and physico-chemical qualities during drying of pulp preconditioned and fermented cocoa (Theobroma cacao) beans. Afr. J. Food Nutr. Sci. 2015;1:9651–9670. [Google Scholar]
  3. Aprotosoaie A.C., Luca S.V., Miron A. Flavor chemistry of cocoa and cocoa products-an overview. Compr. Rev. Food Sci. Food Saf. 2016;15:73–91. doi: 10.1111/1541-4337.12180. [DOI] [PubMed] [Google Scholar]
  4. 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. [DOI] [PubMed] [Google Scholar]
  5. Batista N.N., Ramos C.L., Dias D.R., Pinheiro A.C.M., Schwan R.F. The impact of yeast starter cultures on the microbial communities and volatile compounds in cocoa fermentation and the resulting sensory attributes of chocolate. J. Food Sci. Technol. 2016;53(2):1101–1110. doi: 10.1007/s13197-015-2132-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Beckett S.T. Conching, p. 192–222. In: Beckett S.T., editor. Industrial, Chocolate Manufacture and Use. John Wiley & Sons Ltd; 2009. [Google Scholar]
  7. Bhumibhamon O., Jinda J. Effect of enzymes pectinases on natural cocoa fermentation. Kasetsart J. (Nat. Sci.) 1997;31:206–212. [Google Scholar]
  8. Bonvehí J.S. Investigation of aromatic compounds in roasted cocoa powder. Eur. Food Res. Technol. 2005;221:19–29. [Google Scholar]
  9. Cempaka L., Aliwarga L., Purwo S., Kresnowati M.T.A.P. Dynamics of cocoa bean pulp degradation during cocoa bean fermentation: effects of yeast starter culture addition. J. Math. Fund. Sci. 2014;46:14–25. [Google Scholar]
  10. Crafack M., Mikkelsen M.B., Saerens S., Knudsen M., Blennow A., Lowor S. Influencing cocoa flavour using Pichia kluyveri and Kluyveromyces marxianus in a defined mixed starter culture for cocoa fermentation. Int. J. Food Microbiol. 2013;167:103–116. doi: 10.1016/j.ijfoodmicro.2013.06.024. [DOI] [PubMed] [Google Scholar]
  11. Dalgaard P., Koutsoumanis K. Comparison of maximum specific growth rates and lag times estimated from absorbance and viable count data by different mathematical models. J. Microbiol. Methods. 2001;43 doi: 10.1016/s0167-7012(00)00219-0. www.elsevier.com/locate/jmicmeth Retrieved from: [DOI] [PubMed] [Google Scholar]
  12. Day A., Schneider C., Schneider B.L. Cell Cycle Checkpoint Control Protocols, Book 241. 2004. Yeast cell synchronization; pp. 55–76. [Google Scholar]
  13. De Vuyst L., Weckx S. The cocoa bean fermentation process: from ecosystem analysis to starter culture development. J. Appl. Microbiol. 2016;121:5–17. doi: 10.1111/jam.13045. [DOI] [PubMed] [Google Scholar]
  14. Fisher A.J., Rosenstiel T.N., Shirk M.C., Fall R. Nonradioactive assay for cellular dimethylallyl diphosphate. Anal. Biochem. 2001;292:272–279. doi: 10.1006/abio.2001.5079. [DOI] [PubMed] [Google Scholar]
  15. Gibe A.J.G., Pangan J.O. Evaluation of Pichia mexicana and Bacillus pumilus in hastening fermentation process and its resulting effect on the quality of dried cacao beans. Asian J. Appl. Sci. 2015;3:543–549. [Google Scholar]
  16. Khan F., Latif Z. Molecular characterization of polygalacturonase producing bacterial strains collected from different sources. J. Anim. Plant Sci. 2016;26:612–618. [Google Scholar]
  17. Koné M.K., Guéhi S.T., Durand N., Ban-Koffi L., Berthiot L., Tachon A.F. Contribution of predominant yeasts to the occurrence of aroma compounds during cocoa bean fermentation. Food Res. Int. 2015;89:910–917. [Google Scholar]
  18. Kouame L.M., Doue G.G., Adom N.J., Ouattara H.G., Niamké S.L. Biochemical characterization of microbial populations involved in loh-djiboua cocoa’s fermentation in Cote d’Ivoire. J. Fac. Food Eng. 2015;14:196–205. Ştefan cel Mare University of Suceava, Romania. [Google Scholar]
  19. Krings U., Zelena K., Wu S., Berger R.G. Thin-layer high-vacuum distillation to isolate volatile flavour compounds of cocoa powder. Eur. Food Res. Technol. 2006;223:675–681. [Google Scholar]
  20. Kumar S., Sharma H.K., Sarkar B.C. Effect of substrate and fermentation conditions on pectinase and cellulase production by Aspergillus niger NCIM 548 in submerged (SmF) and solid state fermentation (SSF) Food Sci. Biotechnol. 2011;20:1289–1298. [Google Scholar]
  21. Lagunes Gálvez S., Loiseau G., Paredes J.L., Barel M., Guiraud J.P. Study on the microflora and biochemistry of cocoa fermentation in the Dominican Republic. Int. J. Food Microbiol. 2007;114:124–130. doi: 10.1016/j.ijfoodmicro.2006.10.041. [DOI] [PubMed] [Google Scholar]
  22. Leal G.A., Gomes L.H., Efraim P., De Almeida Tavares F.C., Figueira A. Fermentation of cacao (Theobroma cacao L.) seeds with a hybrid Kluyveromyces marxianus strain improved product quality attributes. FEMS Yeast Res. 2008;8:788–798. doi: 10.1111/j.1567-1364.2008.00405.x. [DOI] [PubMed] [Google Scholar]
  23. Lefeber T., Janssens M., Camu N., De Vuyst L. Kinetic analysis of strains of lactic acid bacteria and acetic acid bacteria in cocoa pulp simulation media toward development of a starter culture for cocoa bean fermentation. Appl. Environ. Microbiol. 2010;76:7708–7716. doi: 10.1128/AEM.01206-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lefeber T., Papalexandratou Z., Gobert W., Camu N., De Vuyst L. On-farm implementation of a starter culture for improved cocoa bean fermentation and its influence on the flavour of chocolates produced thereof. Food Microbiol. 2012;30(2):379–392. doi: 10.1016/j.fm.2011.12.021. [DOI] [PubMed] [Google Scholar]
  25. Lojkowska E., Masciaux C., Boccara W., Robert-baudouy J., Hugouvieux N. Characterization of the pelL gene encoding a novel pectate lyase of Erwinia chrysanthemi 3937. Mol. Microbiol. 1995;16:1183–1195. doi: 10.1111/j.1365-2958.1995.tb02341.x. [DOI] [PubMed] [Google Scholar]
  26. Meersman E., Steensels J., Struyf N., Paulus T., Saels V., Mathawan M. Tuning chocolate flavor through development of thermotolerant Saccharomyces cerevisiae starter cultures with increased acetate ester production. Appl. Environ. Microbiol. 2016;82:732–746. doi: 10.1128/AEM.02556-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Miller G.L. Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal. Chem. 1959;31:426–428. [Google Scholar]
  28. National Institute of Health . 2015. US Library of Medecin.https://hazmap.nlm.nih.gov/category-details?id=481&table=copytblagents [Google Scholar]
  29. 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. [DOI] [PubMed] [Google Scholar]
  30. Ouattara H.G., Koffi B.L., Karou G.T., Sangaré A., Niamke S.L., Diopoh J.K. Implication of Bacillus sp. in the production of pectinolytic enzymes during cocoa fermentation. World J. Microbiol. Biotechnol. 2008;24:1753–1760. [Google Scholar]
  31. Ouattara H.G., Reverchon S., Niamke S.L., Nasser W. Molecular identification and pectate lyase production by Bacillus strains involved in cocoa fermentation. Food Microbiol. 2011;28:1–8. doi: 10.1016/j.fm.2010.07.020. [DOI] [PubMed] [Google Scholar]
  32. Ouattara H.D., Ouattara H.G., Droux M., Reverchon S., Nasser W., Niamke S.L. Lactic acid bacteria involved in cocoa beans fermentation from Ivory Coast: species diversity and citrate lyase production. Int. J. Food Microbiol. 2017;256:11–19. doi: 10.1016/j.ijfoodmicro.2017.05.008. [DOI] [PubMed] [Google Scholar]
  33. Ouattara H.G., Reverchon S., Niamke S.L., Nasser W. Regulation of the synthesis of pulp degrading enzymes in Bacillus isolated from cocoa fermentation. Food Microbiol. 2017;63:255–262. doi: 10.1016/j.fm.2016.12.004. [DOI] [PubMed] [Google Scholar]
  34. Oumer O.J., Abate D. Comparative studies of pectinase production by Bacillus subtilis strain Btk 27 in submerged and solid-state fermentations. BioMed Res. Int. 2018 doi: 10.1155/2018/1514795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Pereira G.V.M., Miguel C P M. G. da, Ramos Cí.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. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Pereira G.V.M., Magalhães K.T., de Almeida E.G., da Silva Coelho I., Schwan R.F. Spontaneous cocoa bean fermentation carried out in a novel-design stainless steel tank: influence on the dynamics of microbial populations and physical-chemical properties. Int. J. Food Microbiol. 2013;161:121–133. doi: 10.1016/j.ijfoodmicro.2012.11.018. [DOI] [PubMed] [Google Scholar]
  37. Pettipher G.L. Analysis of cocoa pulp and the formulation of a standardised artificial cocoa pulp medium. J. Sci. Food Agric. 1986;37:297–309. [Google Scholar]
  38. Qu Y., Feng Y., Wang X., Logan B.E. Use of a coculture to enable current production by Geobacter sulfurreducens. Appl. Environ. Microbiol. 2012;78:3484–3487. doi: 10.1128/AEM.00073-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Ramos C.L., Dias D.R., Miguel C P M. G. da, Schwan R.F. Impact of different cocoa hybrids (Theobroma cacao L.) and S. cerevisiae UFLA CA11 inoculation on microbial communities and volatile compounds of cocoa fermentation. Food Res. Int. 2014;64:908–918. doi: 10.1016/j.foodres.2014.08.033. [DOI] [PubMed] [Google Scholar]
  40. Rikhvanov E.G., Varakina N.N., Sozinov D.Y., Voinikov V.K. Association of bacteria and yeasts in hot springs. Appl. Environ. Microbiol. 1999;65:4292–4293. doi: 10.1128/aem.65.9.4292-4293.1999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Rodriguez-Campos J., Escalona-Buendía H.B., Contreras-Ramos S.M., Orozco-Avila I., Jaramillo-Flores E., Lugo-Cervantes E. Effect of fermentation time and drying temperature on volatile compounds in cocoa. Food Chem. 2012;132:277–288. doi: 10.1016/j.foodchem.2011.10.078. [DOI] [PubMed] [Google Scholar]
  42. Romanens E., Freimüller Leischtfeld S., Volland A., Stevens M., Krähenmann U., Isele D. Screening of lactic acid bacteria and yeast strains to select adapted anti-fungal co-cultures for cocoa bean fermentation. Int. J. Food Microbiol. 2019;290:262–272. doi: 10.1016/j.ijfoodmicro.2018.10.001. [DOI] [PubMed] [Google Scholar]
  43. Samagaci L., Ouattara H.G., Goualié B.G., Niamke S.L. Growth capacity of yeasts potential starter strains under cocoa fermentation stress conditions in Ivory Coast. Emir. J. Food Agric. 2014;26(10):861–870. [Google Scholar]
  44. Samagaci L., Ouattara H., Niamke S., Lemaire M. Pichia kudrazevii and Candida nitrativorans are the most well-adapted and relevant yeast species fermenting cocoa in Agneby-Tiassa, a local Ivorian cocoa producing region. Food Res. Int. 2016;89:773–780. doi: 10.1016/j.foodres.2016.10.007. [DOI] [PubMed] [Google Scholar]
  45. Schwan R.F. Cocoa fermentations conducted with a defined microbial cocktail inoculum. Appl. Environ. Microbiol. 1998;64:1477–1483. doi: 10.1128/aem.64.4.1477-1483.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. 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. [DOI] [PubMed] [Google Scholar]
  47. Schwan R.F., Rose A.H., Board R.G. Microbial fermentation of cocoa beans, with emphasis on enzymatic degradation of the pulp. J. Appl. Bacteriol. 1995;79:96–107. [Google Scholar]
  48. Schwan R.F., Cooper R.M., Wheals A.E. Endopolygalacturonase secretion by Kluyveromyces marxianus and other cocoa pulp-degrading yeasts. Enzym. Microb. Technol. 1997;21(4):234–244. [Google Scholar]
  49. Soumahoro S., Ouattara H.G., Goualié B.G., Koua G., Doue G., Niamke S.L. Occurrence of high acetic acid-producing bacteria in Ivorian cocoa fermentation and analysis of their response to fermentative stress. Am. J. Biosci. 2015;3:70–79. [Google Scholar]
  50. Tantipaibulvut S., Pinisakul A., Rattanachaisit P., Klatin K., Onsriprai B., Boonyaratsiri K. Ethanol production from desizing wastewater using co-culture of Bacillus subtilis and Saccharomyces cerevisiae. Energy Procedia. 2015;79:1001–1007. [Google Scholar]
  51. Van Rijswijck I.M.H., Wolkers - Rooijackers J.C.M., Abee T., Smid E.J. Performance of non-conventional yeasts in co-culture with brewers’ yeast for steering ethanol and aroma production. Microbial Biotechn. 2017;10:1591–1602. doi: 10.1111/1751-7915.12717. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Voigt J., Biehl B. Precursors of the cocoa-specific aroma components are derived from the vicilin-class (7S) globulin of the cocoa seeds by proteolytic processing. Bot. Acta. 1995;108:283–289. [Google Scholar]
  53. Voigt J., Heinrichs H., Voigt G., Biehl B. Cocoa-specific aroma precursors are generated by proteolytic digestion of the vicilin-like globulin of cocoa seeds. Food Chem. 1994;50:177–184. [Google Scholar]
  54. Widdel F. 2010. Theory and Measurement of Bacterial Growth A. Basic and Practical Aspects.https://www.mpi-bremen.de/Binaries/Binary307/Wachstumsversuch.pdf Retrieved from: [Google Scholar]
  55. Yao W., Goualié B.G., Ouattara H.G., Niamké S. Growth capacity of Bacillus potential starter strains isolated from cocoa beans fermentation under culture stress conditions. Sci. Study Res. (Chem. Chemical Eng. Biotechn. Food Ind.) 2017;18(2) [Google Scholar]
  56. Zak O.L., Ostovar K., Keeney P.G. Implication of Bacillus subtilis in the synthesis of tetramethylpyrazine during fermentation of cocoa beans. J. Food Sci. 1972;37(6):967–968. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary materials
mmc1.docx (96.2KB, docx)

Articles from Heliyon are provided here courtesy of Elsevier

RESOURCES