Yeasts are essential for cocoa bean fermentation.

International Journal of Food Microbiology 174 (2014) 72–87

Contents lists available at ScienceDirect

International Journal of Food Microbiology journal homepage: www.elsevier.com/locate/ijfoodmicro

Yeasts are essential for cocoa bean fermentation Van Thi Thuy Ho, Jian Zhao, Graham Fleet ⁎ Food Science and Technology, School of Chemical Engineering, The University of New South Wales, Sydney, New South Wales 2052, Australia

a r t i c l e

i n f o

Article history: Received 4 October 2013 Received in revised form 17 December 2013 Accepted 17 December 2013 Available online 27 December 2013 Keywords: Cocoa bean fermentation Yeasts Lactic acid bacteria Acetic acid bacteria Cocoa quality Natamycin

a b s t r a c t Cocoa beans (Theobroma cacao) are the major raw material for chocolate production and fermentation of the beans is essential for the development of chocolate flavor precursors. In this study, a novel approach was used to determine the role of yeasts in cocoa fermentation and their contribution to chocolate quality. Cocoa bean fermentations were conducted with the addition of 200 ppm Natamycin to inhibit the growth of yeasts, and the resultant microbial ecology and metabolism, bean chemistry and chocolate quality were compared with those of normal (control) fermentations. The yeasts Hanseniaspora guilliermondii, Pichia kudriavzevii and Kluyveromyces marxianus, the lactic acid bacteria Lactobacillus plantarum and Lactobacillus fermentum and the acetic acid bacteria Acetobacter pasteurianus and Gluconobacter frateurii were the major species found in the control fermentation. In fermentations with the presence of Natamycin, the same bacterial species grew but yeast growth was inhibited. Physical and chemical analyses showed that beans fermented without yeasts had increased shell content, lower production of ethanol, higher alcohols and esters throughout fermentation and lesser presence of pyrazines in the roasted product. Quality tests revealed that beans fermented without yeasts were purplish-violet in color and not fully brown, and chocolate prepared from these beans tasted more acid and lacked characteristic chocolate flavor. Beans fermented with yeast growth were fully brown in color and gave chocolate with typical characters which were clearly preferred by sensory panels. Our findings demonstrate that yeast growth and activity were essential for cocoa bean fermentation and the development of chocolate characteristics. Crown Copyright © 2014 Published by Elsevier B.V. All rights reserved.

1. Introduction Cocoa beans (Theobroma cacao) are the major raw material for chocolate production. Fermentation of the beans is essential for removing the pulp that envelops the beans and developing precursors of chocolate flavor (Fowler, 2009; Thompson et al., 2013). Sugars and polysaccharides of the bean pulp are fermented by microorganisms, producing metabolites and conditions that cause bean death and initiate an array of biochemical reactions within the bean that generate chocolate flavor precursors. These flavors are fully developed on subsequent bean roasting and conching as part of the chocolate making process (Afoakwa et al., 2008; Lima et al., 2011; Schwan and Wheals, 2004). Although chocolate manufacturing is a multibillion dollar industry, estimated to have a global value of approximately US$ 95 billion (Pipitone, 2012), bean fermentation is still an uncontrolled traditional process (Pereira et al., 2013) conducted by a consortium of indigenous species of yeasts, lactic acid bacteria and acetic acid bacteria (Schwan and Wheals, 2004; Lima et al., 2011). To transform the fermentation into a more efficient industrialized process, controlled by the use of defined starter cultures, it is essential to know how each microbial group ⁎ Corresponding author at: University of New South Wales, Department of Food Science and Technology, School of Chemical Engineering, Building F10, Room 811, Kensington, New South Wales 2033, Australia. Tel.: +61 2 9385 5664; fax: +61 2 9385 5966. E-mail addresses: [email protected] (V.T.T. Ho), [email protected] (J. Zhao), g.fl[email protected] (G. Fleet).

contributes to the fermentation process and chocolate quality (Saltini et al., 2013). The microbial ecology of cocoa bean fermentation is complex and involves the successional growth of various species of yeasts, lactic acid bacteria, acetic acid bacteria and, possibly, species of Bacillus, other bacteria and filamentous fungi (Ardhana and Fleet, 2003; De Vuyst et al., 2010; Lima et al., 2011; Schwan and Wheals, 2004; Thompson et al., 2013). Within the yeasts, Hanseniaspora guilliermondii or Hanseniaspora opuntiae generally dominate the early part of fermentation after which Saccharomyces cerevisiae, Kluyveromyces marxianus, Pichia membranifaciens, Pichia kudriavzevii and some Candida spp. are most often dominant (Ardhana and Fleet, 2003; Daniel et al., 2009; Galvez et al., 2007; Jespersen et al., 2005; Nielsen et al., 2005). Of the lactic acid bacteria, Lactobacillus plantarum and Lactobacillus fermentum most frequently grow, although contributions from Pediococcus and Leuconostoc species are sometimes reported (Camu et al., 2008b; Kostinek et al., 2008; Nielsen et al., 2007). For the acetic acid bacteria, Acetobacter pasteurianus is most frequently the main contributor, but other species are also involved including Gluconobacter oxydans, Acetobacter tropicalis, Acetobacter lovaniensis and Acetobacter syzygii (Ardhana and Fleet, 2003; Camu et al., 2007; Lefeber et al., 2011; Nielsen et al., 2007). It is not fully understood how these microbial groups or individual species determine cocoa bean quality and chocolate character and, indeed, whether or not they are essential to the fermentation process. Roelofsen (1958) has reviewed early studies on the microbiology of cocoa bean fermentation dating from the late 1890s. Based on these early investigations, some of which involved

0168-1605/$ – see front matter. Crown Copyright © 2014 Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ijfoodmicro.2013.12.014

V.T.T. Ho et al. / International Journal of Food Microbiology 174 (2014) 72–87

crude fermentations with inoculated yeasts and bacteria, it was generally concluded that the growth of yeasts was necessary for successful fermentation and the production of cocoa beans with typical chocolate aromas and character. However, analytical evidence to support this conclusion was lacking. The functional roles of the different groups of microorganisms have been discussed in recent reviews of cocoa bean microbiology by De Vuyst et al. (2010), Lima et al. (2011) and Schwan and Wheals (2004). The yeasts initiate an alcoholic fermentation of the pulp sugars, generating ethanol which, along with acetic acid, enters the bean to kill the embryo and trigger endogenous biochemical reactions that produce the chocolate flavor precursors. The acetic acid is thought to be produced by acetic acid bacteria via oxidation of part of the ethanol produced by the yeasts. The latter reaction generates heat that causes the fermenting bean mass to increase to 45–50 °C, also considered essential for successful fermentation and chocolate flavor development. Some yeasts contribute to pectin degradation in the pulp which facilitates bean aeration and growth of the acetic acid bacteria. The role of lactic acid bacteria and whether or not they are essential to the process is not clear. Their fermentation of pulp sugars to give lactic acid can be detrimental to cocoa bean and chocolate quality, leading to too much acidity. However, their production of this acid and potential to utilize the citric acid of the pulp may contribute to the pH balance of the process. It is believed that the internal pH of cocoa bean (around 7.0 before fermentation) must decrease to between pH 5.0 and 5.5 to allow good activity of endogenous proteases that are essential to the degradation of bean proteins and production of chocolate flavor precursors (Biehl et al., 1985; Hansen et al., 1998). Some early studies have suggested an important role of heterofermentative lactic acid bacteria in producing acetic acid, necessary for killing the beans (Roelofsen, 1958). To better understand how individual microbial groups and species contribute to cocoa bean fermentation and chocolate character, controlled fermentation studies that integrate microbiological, chemical and sensory analyzes are required. The objective of this paper is to determine the contribution of yeasts, as a group, to cocoa bean fermentation and chocolate character. To investigate their contribution, we conducted controlled cocoa bean fermentations in the presence and absence of Natamycin (Natamax, Danisco, Denmark) which is an approved food additive that inhibits yeast growth (European Food Safety Authority, 2009).

73

were mixed every 48 h. The fermentations were stopped at day 6 when the beans were removed from their boxes and dried as a single layer at 30 °C and RH 70% for 5 days. Samples of beans (100 g total, from locations throughout the fermenting mass) were taken daily for microbiological and chemical analyses. Samples for microbiological analysis were used immediately while those for chemical analysis were stored at −20 °C until examined. The fermentations were repeated twice, once using cocoa beans harvested in March 2011 and again with beans harvested from another plantation in October 2011. Several preliminary fermentations were conducted to optimize the conditions for fermentation and drying, chocolate production and sensory assessment, and to select the appropriate concentration of Natamycin. 2.2. Microbiological analyses 2.2.1. Enumeration of microbial populations Cocoa beans (25 g) were aseptically mixed with 225 ml of 0.1% peptone water in a Stomacher bag and manually shaken for 5 min to give a uniform suspension of the pulp material. One ml of the suspension was serially diluted in 0.1% peptone water and 0.1 ml samples from each of three consecutive dilutions were spread inoculated onto duplicate plates of different agar media. The enumeration of yeasts was done on plates of Malt Extract Agar (MEA) (Oxoid) containing 100 mg/l of oxytetracycline and Dichloran Rose Bengal Chloramphenicol Agar (DRBC) with incubation at 25 °C for 3–4 days. Lactic acid bacteria were enumerated on de Man Rogosa Sharpe (MRS) Agar (Oxoid) containing 100 mg/l of cycloheximide at 30 °C for 3–4 days. Acetic acid bacteria were enumerated on Wallerstein Laboratories Nutrient Agar (WLNA) (Oxoid) and Glucose–Yeast Extract Agar (GYEA) (5% glucose, 0.5% yeast extract and 1.5% agar) containing 100 mg/l of cycloheximide at 30 °C for 4–5 days. Plates for culture of yeasts and acetic acid bacteria were incubated under aerobic conditions while a candle-jar was used for the incubation of lactic acid bacteria. After incubation, counts of the yeast and bacterial groups were determined as well as those of individual yeast and bacterial species following observations of their colonial and cellular morphologies and subsequent identification. At least three representatives of each colony type were isolated from each sampling time, and purified for identification. Population data reported are the means of duplicate analyses.

2. Materials and methods 2.1. Cocoa bean fermentation Cocoa pods (Trinitario variety) were harvested from plantations in North Queensland, Australia and transported to the University of New South Wales, Sydney. The pods were stored at 20–25 °C for 7–10 days, then cut with a knife for manual removal of the beans. The mass of beans was uniformly mixed on a tray and deliberately brought into contact with the outer surfaces of extracted pods to provide a source of natural microbial inoculum. The bean mass was divided into two × 5 kg batches which were transferred into two plastic fermentation boxes (17 × 17 × 20 cm) with drilled holes on the sides and the base to facilitate juice drainage and aeration. A solution of Natamycin was sprayed onto the beans in one box, to inhibit yeast growth. The beans were mixed to give a final concentration of 200 ppm of Natamycin throughout the mass. The cocoa beans in the other box were not treated with Natamycin but similarly mixed. The boxes of beans were incubated for fermentation which developed spontaneously due to growth of the indigenous microflora. For fermentation, the boxes were covered with lids and incubated at 25 °C (0–12 h), 30 °C (12–24 h), 35 °C (24–36 h), 40 °C (36–48 h), 45 °C (48–72 h) and 48 °C (72–144 h) to simulate the temperature evolution that occurs during commercial cocoa fermentations (Schwan and Wheals, 2004; Lima et al., 2011). The fermenting beans

2.2.2. Identification of yeast and bacterial species Cellular morphologies of the isolates were determined by phase contrast microscopic observation. For the bacteria, Gram stains and catalase tests were conducted. Yeasts and bacteria were identified to species by a combination of rDNA sequencing and restriction fragment length polymorphism (RFLP) analysis. 2.2.2.1. DNA extraction and amplification by Polymerase Chain Reaction (PCR). Extraction of DNA from cells of yeasts (grown in Malt Extract Broth) and bacteria (grown in MRS broth) followed the protocol of Cocolin et al. (2002). The 5.8S-Internally Transcribed Spacer (5.8S-ITS) rDNA region of yeast isolates was amplified by PCR using the primers ITS1 (5′-TCCG TAGGTGAACCTGCGG-3′) and ITS4 (5′-TCCTCCGCTTATTGATATGC-3′) as described by Esteve-Zarzoso et al. (1999). The 16S rDNA of bacterial isolates was amplified by PCR using the primers 27 F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1495R (5′-CTAC GGCTACCTTGTTACGA-3′) for lactic acid bacteria (Yu et al., 2009) and the primers 16Sd (5′-GCTGGCGGCATGCTTAAC ACAT-3′) and 16Sr (5′GGAGGTGATCCAGCCGCAGGT-3′) for acetic acid bacteria (Ruiz et al., 2000). The primers were purchased from SigmaAldrich (Sydney, NSW, Australia) and the nucleotides and enzyme from New England Biolabs (Ipswich, MA, USA).

74


The purity of PCR products was checked by electrophoresis in 1% agarose gel followed by staining with GelRed dye (Biotium, Hayward, CA, USA) and used for RFLP and sequence analyses. 2.2.2.2. RFLP analysis. For RFLP analysis, the following restriction enzymes (New England Biolabs) were used CfoI, HaeIII and HinfI for yeasts and HinfI, HaeIII, TaqI and RsaI for lactic acid and acetic acid bacteria. The amplified DNAs (5 μl) were digested by 5 U of restriction enzyme in 20 μl reaction mixtures for 2 h at 37 °C, except for the TaqI reaction which was conducted at 65 °C. Restriction fragments were analyzed by horizontal 3% agarose gel electrophoresis. DNA ladder markers of 20 and 100 bp (GeneWorks, Adelaide, Australia) were used as the size standard. Electrophoresis was performed at 140 V for 3 h and the gels were stained with GelRed dye and photographed under transluminated UV light. To determine species identity, the restriction fragment profiles were compared with those of reference cultures and those published by Esteve-Zarzoso et al. (1999) and Granchi et al. (1999) for yeasts, Yu et al. (2009) for lactic acid bacteria and Ruiz et al. (2000) for acetic acid bacteria. The following reference cultures were obtained from the collection of the School of Biotechnology and Biomolecular Sciences, the University of New South Wales: S. cerevisiae (No. 508100), P. kudriavzevii (No. 406600), K. marxianus (No. 509000), Hanseniaspora uvarum (No. 502900), L. plantarum (No. 026900), L. fermentum (No. 053300), A. pasteurianus (No. 092200) and G. oxydans (No. 030300). 2.2.2.3. Sequence analysis. PCR products (50–100 ng) were purified using ExoSAP-IT clean up enzyme (USB Corporation, Cleveland, OH, USA). The purified PCR product was mixed with 9.6 pmol of respective forward or reverse primers of each microbial group as described in Section 2.2.2.1 and sent to the Australian Genome Research Facility (Westmead Millennium Institute, Sydney, Australia) for sequencing. The sequences of the 5.8S-ITS rDNA region in yeasts and the 16S rDNA in bacteria were subjected to online reference data at the GenBank database using the Basic Local Alignment Search Tool and identified by comparing sequence homology with available microbial data. 2.3. Chemical analyses 2.3.1. pH measurement The measurement of pH of cocoa beans followed the protocol of Senanayake et al. (1997). For pulp pH, 10 whole cocoa beans with attached pulp were shaken in 100 ml of MiliQ water for 15 min. The beans were separated by decanting and the pH of the supernatant was measured using a digital pH meter. For nib pH, the shells of the whole beans were peeled to obtain the nibs (cotyledons) which were then ground in a BCG300 grinder (Breville, Sydney, Australia). Aliquots (10 g) of the ground sample were stirred in 100 ml of MiliQ water for 15 min and the pH of the supernatant was measured as described previously. 2.3.2. Sugars, ethanol, glycerol, mannitol and organic acids Cocoa pulp and nib fractions were analyzed for their concentrations of sugars, ethanol, glycerol, mannitol and organic acids by high performance liquid chromatography (HPLC). The pulp and nibs from whole cocoa beans were manually separated and ground. Samples (5 g) of each fraction were mixed with 60 ml of MilliQ water and homogenized in a Waring blender (John Morris Scientific, Sydney, Australia) for 3 min. The homogenate was centrifuged at 15,000 rpm for 20 min at 5 °C and the supernatant was retained. The sediment was washed twice with 20 ml of MilliQ water and all the supernatants were pooled and clarified by filtering through 0.45 μm Hydraflon™ membranes (MicroAnalytix, Sydney, Australia) before analysis by HPLC. The instrumentation for HPLC was a Shimadzu system comprising a LC-20 AD pump, a SIL-20A HT auto injector, a RID-10A detector, a SPDM20A Photodiode Array detector and a column heater and a Rezex ion-

exclusion ROA column (Phenomenex, Torrance, CA, USA). The mobile phase of 0.01 M H2SO4 at a flow rate of 0.3 ml/min was used to elute sugars, ethanol, glycerol and mannitol at 30 °C and organic acids at 45 °C. Sugars, ethanol, glycerol and mannitol were detected by the refractive index detector and organic acids by the Photodiode Array detector at 210 nm. The data were recorded and analyzed by the LC Solution software of the instrument. The concentrations of individual sugars, ethanol, glycerol, mannitol and organic acids were determined by comparison with standard curves constructed from standard solutions of sucrose, glucose, fructose, ethanol, glycerol and mannitol at 20, 10, 5 and 1 mg/ml and oxalic, citric, malic, succinic, lactic and acetic acids at 2, 1, 0.5 and 0.1 mg/ml for each compound. Analyses of individual samples were done in duplicate and average values are reported. 2.3.3. Volatile compounds Volatile compounds in separate fractions of cocoa pulp and nibs were extracted by head space–solid phase microextraction (HS-SPME) using a 50/30 μm divinylbenzene/carboxene/polydimethylsiloxane fiber (Supelco, Bellefonte, PA, USA). Ground cocoa sample (1 g) was put into a 4 ml headspace vial and sealed with a septum cap. The extraction conditions and analysis were adapted from the method of Rodriguez-Campos et al. (2011): 15 min for reaching equilibrium and 30 min for exposing the fiber to the volatile compounds in the HS at 60 °C. The extracted cocoa volatiles were analyzed using an Agilent 5975 gas chromatography–mass spectrometer (GC-MS) equipped with a Hewlett Packard capillary column (HP-5MS, 30 m length × 0.25 mm ID × 0.25 μm film thickness). The GC oven temperature was initially set at 30 °C for 5 min, increased to 200 °C at a rate of 3 °C/min and then increased at a rate of 5 °C/min to 250 °C which was maintained for 5 min. The carrier gas was helium. Injection mode was splitless at 250 °C for 1 min. The mass spectrometer was set as follows: mass-ionization at 70 eV; ion-source temperature at 280 °C and data were recorded by the Chemstation software. Identification of volatile compounds was obtained by comparing (1) the mass spectra of individual compounds with the Wiley library, (2) the retention time with literature data using the Kovac index and (3) the mass spectra of analyzed compounds with those of pure standards which were hexanal, nonanal, linalool, isoamyl alcohol, ethyl acetate, ethyl phenylacetate, ethyl isovalerate, 2-pentyl acetate, ethyl benzoate, 2-heptanone, 2-nonanone, acetophenone, toluence, ethylbenzene, decane, styrene and 2,3,5 trimethyl pyrazine. The volatile compounds in each sample were extracted and analyzed in duplicate and results were expressed as an average. 2.4. Free amino acids Aqueous extracts of cocoa nibs prepared from beans fermented for 0, 72 and 144 h as described in Section 2.3.2 were analyzed at the Australian Proteome Analysis Facility Ltd (Macquarie University, Sydney, Australia). The analysis details included (1) mixing equal volumes of sample and an internal standard (Norvaline), (2) filtering the samples through a 10 k DA molecular weight cut off spin filter and (3) analyzing the filtrates by the Waters AccQTag chemistry on a Waters Acquity UPLC system. The samples were analyzed in duplicate and average values are reported. 2.5. Alkaloid and polyphenolic compounds The defatted cocoa nib fractions (5 g) were analyzed for their concentrations of theobromine, caffeine, (+)-catechin and (−)-epicatechin according to the extraction and HPLC methods of Cooper et al. (2007) and using standard solutions of theobromine, caffeine, (+)-catechin and (−)-epicatechin (SigmaAldrich, Sydney, NSW, Australia) at 0.01,


0.05. 0.1 and 0.2 mg/ml. The samples were analyzed in duplicate and the results were expressed as an average.

75

3. Results 3.1. Microbial ecology of cocoa bean fermentation

2.6. Sodium Dodecyl Sulfate (SDS)–Polyacrylamide Gel Electrophoresis (PAGE) SDS–PAGE was used to examine the degradation of cocoa cotyledon proteins during fermentation. The methods for extracting proteins and running electrophoresis followed Buyukpamukcu et al. (2001).

2.7. Quality evaluation of cocoa beans 2.7.1. Cut test The cut test was used to compare the quality of dried cocoa beans. Briefly, 30 dried, fermented cocoa beans were cut lengthwise from the middle using a sharp knife so that the largest surface of cotyledon was exposed. Each half of the beans was visually inspected and classified based on their color (fully brown, partly brown, partly purple, purple or slaty) and defects (germinated or insect infested) (Wood and Lass, 1985).

2.7.2. Shell content To determine shell content, 20 dried, fermented beans were weighed and the shell manually removed from the nibs. The shell content was calculated as the percentage of shell weight over the whole bean weight. Analyses were done in duplicate and average values are reported.

2.7.3. Chocolate making and sensory evaluation Sensory evaluations were conducted on chocolates prepared from three separate batches of fermented beans. These batches were from the March 2011 and October 2011 beans mentioned in Section 2.1 and a third batch that was harvested in August 2010 and fermented in the presence and absence of Natamycin. Dried, fermented beans were roasted at 120 °C for 55 min. Cocoa nibs were obtained by manually removing the shells of roasted beans. Deshelled nibs (111 g) were ground into cocoa liquor and transferred to a laboratory scale conching machine, Spectra 11 (Santha Industrials, Mangalore, Karnataka, India). After an initial conching for 10 min, 72 g of pure icing sugar (Sugar Australia, Melbourne, Australia) and 45 g of cocoa butter (ADM, Decatur, IL, USA) were added. The mixture was conched for 6 h until the liquid chocolate became fine. The temperature of the chocolate increased from 30 °C to about 40 °C throughout the process. The liquid chocolate was transferred to a bowl for tempering by heating to 45 °C, cooling to 27 °C and finally bringing it to 33 °C. The chocolate was poured into paper molds, covered and stored in an odor free refrigerator for at least 2 weeks before being tasted. Sensory evaluation of chocolate was performed by an untrained panel of 90 chocolate consumers (30 for each batch of chocolate) aged between 20 and 65, using the liking test method as described by Carpenter et al. (2000). The panelists were asked to compare and evaluate color, chocolate flavor and overall acceptability of two coded samples of chocolate by giving scores in a seven-point hedonic scale corresponding to their liking.

2.8. Statistical analysis One-way single factor analysis of variance and t-test were used to determined significant differences between means using Microsoft Excel. Significant differences in the concentrations of volatile compounds, shell content and sensory evaluation of chocolates were considered when p b 0.05.

Fermentation samples were examined for yeasts and bacteria by simultaneous cultivation on different media to increase the reliability of data. In the case of yeasts, similar total populations were found on plates of either MEA or DRBC agar. Both media gave isolation of the same species of yeasts at similar populations. For acetic acid bacteria, similar populations and species were obtained on GYEA and WLNA. Although MRS agar was used for the enumeration and isolation of lactic acid bacteria, the two main species of acetic acid bacteria found in this study (Gluconobacter frateurii and A. pasteurianus) grew well on this medium and were easily distinguished by their colony morphologies. WLNA also supported the growth of Pantoea agglomerans that was sometimes found at the commencement of fermentations, and Bacillus species found in the late stages of some fermentations. Cocoa beans consistently underwent spontaneous fermentation due to the growth of naturally associated yeasts and bacteria. Preliminary experiments were conducted to determine if Natamycin (50, 100 and 200 ppm) would prevent the growth of yeasts associated with these fermentations. Fermentations with 100 and 200 ppm of Natamycin showed no yeast growth after 6–8 days, whereas yeasts were not entirely suppressed in fermentations with 50 ppm of Natamycin. Changes in the pH of the cocoa bean pulp and nibs were similar for fermentations with 100 and 200 ppm of Natamycin. The growth profiles of total bacteria were similar for fermentations in the presence of either 100 or 200 ppm of Natamycin. It was decided to use 200 ppm of Natamycin in subsequent fermentations. Fig. 1 shows the growth of total yeasts and bacteria during the fermentation of cocoa beans in the absence and presence of Natamycin for beans harvested in March 2011 (Fig. 1a, c) and October 2011 (Fig. 1b, d). Both yeasts and bacteria (approximately 103–104 cfu/g) were present at the beginning of fermentations where no Natamycin had been added and they subsequently grew to maximum populations of 106–108 cfu/g for the yeasts and 108–109 cfu/g for the bacteria within 48–72 h (Fig. 1a, b). After that, their populations decreased, this being greater for the October beans. No yeasts were detected in fermentations containing Natamycin (Fig. 1c, d). However, the growth of bacteria was unaffected by Natamycin, giving a similar profile to that obtained in the absence of Natamycin. For both the March and October trials, there were no statistically significant differences (p N 0.05) for the growth of total bacteria in fermentations conducted in the presence or absence of Natamycin. Thus, while Natamycin inhibited the yeasts, it did not affect the growth of the bacteria during fermentation. Three yeast species dominated the fermentations with no Natamycin and grew sequentially. Similar trends were found for the beans fermented in March or October (Fig. 2a, b). H. guilliermondii (Accession No. FJ491945.1) was found at the commencement of fermentation at 102–103 cfu/g, grew to 106–107 cfu/g by 48 h, and then decreased to undetectable levels (b102 cfu/g). P. kudriavzevii (formerly Issatchenkia orientalis; Accession No. EU798698.1) grew to maximum populations of 106–107 cfu/g at 72–96 h before declining. K. marxianus (Accession No. DQ249190.1) was not detected (b 102 cfu/g) until 72 h of fermentation, after which it grew to 105–106 cfu/g. Its growth kinetics were different for the March and October fermentations, developing earlier in the latter fermentation and dying off by the end of fermentation at 144 h. Pichia guilliermondii (Accession No. EF197951.1), Candida intermedia (Accession No. DQ646683.1) and Cryptococcus flavescens (Accession No. AM160631.1) were detected at 102–104 cfu/g during the first 24 h of the fermentations and declined to undetectable levels (b102 cfu/g) by 48 h. The same species of bacteria and similar profiles of growth were found for fermentations in the absence or presence of Natamycin (Fig. 2). This conclusion was evident for both the March and October fermentations, but there were some differences in the species found for these two trials.

76


Fig. 1. Changes in the populations of total yeasts (●) and total bacteria (■) during untreated (a, b) and Natamycin treated (c, d) fermentations of cocoa beans harvested in March 2011 (a, c) and in October (b, d).

In the March trial, L. plantarum (Accession No. HQ117897.1) and two species of acetic acid bacteria (G. frateurii, Accession No. JF794021.1 and A. pasteurianus, Accession No. GQ240639.1) dominated the fermentations

(Fig. 2c, e). G. frateurii was detected at the start of fermentations and reached maximum populations of 107–108 cfu/g by 48 h, after which it decreased to about 102 cfu/g by 96 h and 120 h for fermentations in

Fig. 2. The growth of yeast and bacterial species during fermentations of cocoa beans harvested in March 2011 (a, c, e) and in October 2011 (b, d, f).


the absence and presence of Natamycin, respectively. A. pasteurianus was not detected in the fermentation without Natamycin until 24 h and in the fermentation with Natamycin until 48 h. It grew to a maximum level of about 106 cfu/g by 72 h in both fermentations, thereafter decreasing to less than 102 cfu/g. L. plantarum was first detected in both fermentations at 24 h after which it grew to 106–107 cfu/g (72–96 h) before declining (Fig. 2c, e). It was the main bacterial species detected at the end of fermentation (144 h). Pantoea agglomerans (Accession No.DQ855292.1) and Asaia sp. (Accession No.FN297840.1) were isolated from the unfermented beans at 102–103 cfu/g, but their populations decreased to undetectable levels (b 102 cfu/g) by 24 h of fermentation (data not shown). For the fermentations conducted in October 2011, there was a weaker presence of G. frateurii and no A. pasteurianus was detected (Fig. 2d, f). In addition to L. plantarum, two other species of lactic acid bacteria, Lactococcus lactis and L. fermentum were found. The kinetics of their growth were similar for fermentations in the absence and presence of Natamycin. L. fermentum was the dominant species in the later stages of fermentation. Bacillus subtilis started to grow during the fermentation of these beans after 96 h (Fig. 2d, f).

3.2. Chemical changes during cocoa bean fermentation 3.2.1. pH changes in cocoa pulp and nibs Fig. 3 shows changes in the pH of the cocoa pulp and nibs during fermentation of the beans. Before fermentation, the pH of the cocoa pulp from beans harvested in March was 3.8. During fermentation in the absence or presence of Natamycin, the pH of the pulp decreased to about 3.0 in the first 48 h, after which it increased to 3.6 for fermentation in the absence of Natamycin, but only to 3.2 when Natamycin was present (Fig. 3a). A similar trend was found for the beans harvested in October, but the initial pH of the pulp was 4.1 and the final pH was 4.2 in the absence of Natamycin and 3.9 in the presence of Natamycin (Fig. 3b). For the nibs of March beans, the initial pH was 6.8 which decreased during fermentation to a final value of 5.4 for fermentation in the absence of Natamycin and 5.0 in the presence of Natamycin (Fig. 3c). The unfermented nibs of the October beans had a pH of 6.5 and this decreased to 4.5 for beans fermented in the absence of Natamycin and 4.3 for those fermented with Natamycin (Fig. 3d). In two other trials (data

77

not shown), nib pH at the end of fermentations was always lower by 0.2-0.4 units in fermentations where yeast growth was suppressed.

3.2.2. Changes in the concentration of sugars, ethanol, glycerol, mannitol and organic acids Fig. 4 shows changes in the concentrations of sugars and ethanol in the pulp and nib fractions during fermentation of cocoa beans. Their concentrations were calculated on both a wet weight and dry weight basis, but only the wet weight results are presented to facilitate comparison with data in the literature. Data for the fermentations conducted in March 2011 are presented. Similar trends and conclusions were obtained with the beans fermented in October 2011 and are not presented, but relevant data will be mentioned where major differences were observed. In unfermented cocoa beans, the main sugars in the pulp were glucose (48 mg/g) and fructose (71 mg/g), while those in the nibs were low levels of sucrose (about 17 mg/g), glucose (0.5 mg/g) and fructose (1 mg/g). During fermentation in the absence of Natamycin, the concentrations of glucose and fructose in the pulp decreased with fructose completely utilized and glucose reduced to 17.5 mg/g (Fig. 4a). For beans fermented in the presence of Natamycin, the concentrations of the two sugars in the pulp decreased at a slower rate, giving considerably higher residuals of fructose (16 mg/g) and glucose (35 mg/g) (Fig. 4b). However, changes in the concentrations of sugars in the nibs followed similar trends for fermentations in the absence or presence of Natamycin. Sucrose concentration in the nib dropped to undetectable levels in both fermentations and the low initial levels of glucose and fructose increased to about 4 mg/g (Fig. 4c, d). Ethanol was not detected in the pulp or nibs of cocoa beans before fermentation. For fermentation in the absence of Natamycin, low concentrations of ethanol were detected in the pulp (0.4 mg/g) and in the nib (0.2 mg/g) at 24 h, which increased rapidly to a maximum of 6.5 mg/g in the pulp and 5.2 mg/g in the nib by 48 h. Thereafter, the concentrations of ethanol decreased, giving residual amounts of 2.3 mg/g and 2.4 mg/g in the pulp and nib, respectively, at the end of the process (Fig. 4e). In contrast, production of ethanol in the Natamycin-added fermentation was very low, with maximum concentrations of 0.3 mg/g in the pulp and 0.2 mg/g in the nib (Fig. 4f).

Fig. 3. Changes in the pH of the pulp (a, b) and nibs (c, d) during untreated (●) and Natamycin treated (■) fermentations of cocoa beans harvested in March 2011 (a, c) and in October 2011 (b, d).

78


Fig. 4. Changes in the concentrations of sucrose (▲), glucose (●) and fructose (■) in the pulp (a, b) and nibs (c, d); and changes in the concentrations of ethanol (e, f) in the pulp (○) and in the nibs (□) during untreated (a, c, e) and Natamycin treated cocoa bean fermentations (b, d, f).

Glycerol was not present in the pulp or nibs of the unfermented cocoa beans. It was detected in the pulp of beans fermented in the absence of Natamycin by 48 h at 0.4 mg/g (data not shown). It increased gradually to maximum levels of 2.5 mg/g at 120–144 h. Glycerol was not detectable in the nibs of beans fermented in the absence of Natamycin. For fermentation in the presence of Natamycin, glycerol was not found in either pulp or nibs. Although the general conclusions for sugars, glycerol and ethanol were similar for beans fermented in March and October, there were some notable differences. Mannitol was found in the pulp and nib fractions of beans for the October fermentation but not the March fermentations. The production of mannitol was first detected in these pulps (3.2–4 mg/g) at 48 h and in the nibs (1–1.2 mg/g) at 72 h. It increased to maximum concentrations of 32 mg/g in the pulp and 5 mg/g in the nib by the end of the fermentation without Natamycin, and to 21 mg/g and 4 mg/g respectively, for fermentations in the presence of Natamycin. Its production coincided with the growth of L. fermentum (Fig. 2d, f). With regard to ethanol, greater maximum concentrations (12 mg/g in the pulp and 8 mg/g in the nibs) and final concentrations (5.6 mg/g in the pulp and 4.2 mg/g in the nibs) were produced for fermentations in the absence of Natamycin compared with those given in Fig. 4e. For fermentations in the presence of Natamycin, very little ethanol was produced and

its maximum concentrations in the pulp or nibs were less than 1 mg/g. Fig. 5 shows changes in concentrations of organic acids on a dry weight basis during cocoa fermentation. In the pulp of unfermented cocoa beans, citric acid (44–45 mg/g) was the main acid while malic (7–8 mg/g) and succinic (5–6 mg/g) acids were also detected (Fig. 5a, b). Changes in the pulp concentrations of citric and succinic acids throughout fermentation were similar for both the Natamycin treated and untreated fermentations, although a small decrease in the concentration of citric acid was noted in the initial stages of fermentation in the absence of Natamycin. By the end of both fermentations, the concentrations of citric and succinic acids had increased to 50–60 mg/g, and 15–20 mg/g respectively. Malic acid had decreased to 3 mg/g in the fermentation without Natamcyin while it remained relatively unchanged in the fermentation with Natamycin. Lactic and acetic acids were not found in the pulp of fresh cocoa beans. Lactic acid was first detected in the pulp at 24 h and its kinetics of production to maximum values about 25–29 mg/g were similar for both fermentations (Fig. 5a, b). However, the kinetics of acetic acid production were different for the two fermentations. The production of acetic acid was faster in the fermentation without Natamcyin than in the one with Natamycin. Acetic acid was first detected by 48 h at 12.5 mg/g in the fermentation without Natamcyin while only at 2 mg/g in the Natamcyin treated


79

Fig. 5. Changes in the concentrations of organic acids in the pulp (a, b) and nibs (c, d) during untreated (a, c) and Natamycin treated cocoa bean fermentations (b, d); citric acid (■), succinic acid (▲), lactic acid (○), acetic acid (●), malic acid (□) and oxalic acid (Δ).

fermentation. Although acetic acid reached maximum values of 14–15 mg/g in both fermentations by 72 h, its subsequent decline was less notable in the Natamycin treated fermentation (Fig. 5a, b). Unfermented cocoa nibs contained citric (16–17 mg/g), succinic (13–14 mg/g), malic (1–2 mg/g) and oxalic (7–8 mg/g) acids. During both fermentations, the concentration of malic and oxalic acids remained relatively unchanged while that of succinic acid showed a slight increase in the late phase of the process (Fig. 5c, d). The level of citric acid gradually decreased during the first 96 h but after that it increased towards the end of fermentations. No lactic or acetic acids were found in unfermented nibs, but their concentrations increased as fermentation progressed. Nib lactic acid was found at 2–3 mg/g after 48 h of fermentation, and increased progressively to 10–13 mg/g at the end for both fermentations. Acetic acid was detected earlier (48 h) in the nibs from fermentations in the absence of Natamycin compared with those fermented with Natamycin (72 h). For both fermentations, the concentration of nib acetic acid increased to maximum levels of about 16–20 mg/g at 96 h and subsequently decreased to 9–10 mg/g by the end of the fermentations (Fig. 5c, d). The beans fermented in October gave similar profiles to those of the March beans for the behavior of malic, lactic, acetic and succinic acids. However, in the October fermentations where Lactococcus lactis and L. fermentum grew, the concentrations of citric acid in the pulp reduced remarkably by about 78% throughout the process (data not shown). 3.2.3. Production of volatile compounds during fermentation The SPME-GC/MS method for volatile detection and measurement enabled the identification of at least 40 compounds in most samples. These compounds were grouped into aldehydes, ketones, esters, alcohols, pyrazines, terpenes and others. Fig. 6 shows the concentrations of the groups of volatile compounds in the cocoa pulp and nibs during fermentation of the March 2011 beans. Fig. 7 shows changes in the concentration of some key volatiles in the pulp and nibs. Similar trends were found for the beans fermented in October 2011 (data not reported), with major differences mentioned in the text. 3.2.3.1. Higher alcohols. Very little higher alcohol was detected in the pulp of unfermented beans. However, this concentration increased about 25 fold during fermentation in the absence of Natamycin but did not increase for fermentation in the presence of Natamycin

(Fig. 6a). Three main higher alcohols were found and these were phenylethyl alcohol (81.5% of total), isoamyl alcohol (10.8%) and 2methyl-1-butanol (4.2%). The data in parenthesis indicate their relative amounts at the end of fermentation in the absence of Natamycin. The concentration of higher alcohols in the nibs before fermentation was almost non-detectable. However, this low amount increased 20 fold throughout fermentation in the absence of Natamycin and then decreased about 2 fold during drying and roasting (Fig. 6b). Phenylethyl alcohol (64%), 2-methyl-1-butanol (23%) and isoamyl alcohol (12%) were the three main alcohols detected in the nibs. The data in parenthesis show the relative amounts in roasted beans fermented in the absence of Natamycin. There was little increase (statistically not significant) in higher alcohol concentration in the nibs fermented in the presence of Natamycin (Fig. 6b). The kinetics of phenylethyl alcohol and isoamyl alcohol production in the pulp and movement into the nibs are shown in Fig. 7a and b. 3.2.3.2. Esters. Total ester concentration was almost non-detectable in unfermented pulps but this concentration increased about 74 fold during fermentation in the absence of Natamycin but less than 3 fold in the presence of Natamycin (Fig. 6c). Five main esters were found in fermenting pulps and their relative amounts at the end of fermentation without Natamycin were phenethyl acetate (49.7%), ethyl phenyl acetate (19.7%), ethyl hexanoate (10%), ethyl octanoate (8.7%) and isoamyl acetate (7.5%). Total ester concentration in the nibs increased approximately 45 fold in fermentation without Natamycin and 15 fold in fermentation with Natamycin (Fig. 6d). The ester concentrations in the nibs from both fermentations decreased after roasting but significant amounts remained in the roasted beans. The main esters in roasted nibs from fermentation in the absence of Natamycin were isoamyl acetate (35%), phenethyl acetate (27.5%) and 2-methyl-butyl acetate (23%). Four other esters found in the nibs but not in the pulp were ethyl isovalerate, 2-methyl-butyl acetate, n-hexyl acetate and ethyl benzoate. Fig. 7c and d show changes in the concentration of isoamyl acetate and phenethyl acetate in the pulp and nibs during fermentation and the continued development of much higher concentrations in the nib during the drying stage. Ethyl acetate was not found in the pulp or nibs of the beans fermented in March 2011. However, it was the main ester along with isoamyl acetate in the October beans, being produced in the pulp and transferring to the nibs during fermentation. Very little production of

a) Higher alcohols

1.5E+9 1.2E+9 9.0E+8 6.0E+8 3.0E+8 0.0E+0 0

24

48

72

96

120 144

Area unit/g nibs (dry basis)


Area unit/ g pulp (dry basis)

80

b) Nib alcohols

1.5E+9 1.2E+9 9.0E+8 6.0E+8 3.0E+8 0.0E+0 0

c) Pulp esters

3.5E+8 3.0E+8 2.5E+8 2.0E+8 1.5E+8 1.0E+8 5.0E+7 0.0E+0 0

24

48

72

96

120 144

3.0E+8 2.5E+8 2.0E+8 1.5E+8 1.0E+8 5.0E+7 0.0E+0 0

2.0E+9 1.5E+9 1.0E+9 5.0E+8 0.0E+0 0

24

48

72

96

2.0E+9 1.5E+9 1.0E+9 5.0E+8 0.0E+0

120 144

0

8.0E+7 6.0E+7 4.0E+7 2.0E+7 0.0E+0 0

24

48

72

96

120 144

Fermentation time (h)

24 48 72 96 120 144 D

R



Area unit/g pulp (dry basis)

g) Pulp ketones

R

f) Nib aldehydes

2.5E+9

Fermentation time (h) 1.0E+8

24 48 72 96 120 144 D

Fermentation time (h) Area unit/g nibs (dry basis)


e) Pulp aldehydes

R

d) Nib esters

3.5E+8

Fermentation time (h) 2.5E+9

24 48 72 96 120 144 D

Fermentation time (h) Area unit/g nibs (dry basis)



h) Nib ketones

1.0E+8 8.0E+7 6.0E+7 4.0E+7 2.0E+7 0.0E+0 0

24 48 72 96 120 144 D

R


Fig. 6. Total concentrations of alcohols (a, b), esters (c, d), aldehydes (e, f) and ketones (g, h) produced in the cocoa pulp (a, c, e, g) and nibs (b, d, f, h) during fermentation, drying (D) and roasting (R) in untreated ( ) and Natamycin-treated fermentations (■).

these esters occurred during fermentation in the presence of Natamycin. 3.2.3.3. Aldehydes. Unfermented pulp and nibs contained detectable levels of aldehydes, with higher levels being found in the nibs. For both fractions, the levels increased throughout fermentation but there were no statistically significant differences (p N 0.05) for fermentations in the absence or presence of Natamycin (Fig. 6e, f). Four main aldehydes were detected in the pulp and nibs. These were phenyl

acetaldehyde (47%, 46%), benzaldehyde (31.3%, 31%), nonanal (15%, 18%) and haxanal (6%, 4.8%), the relative values given being those in the pulps at the end of fermentation in the absence and presence of Natamycin, respectively. In the case of the nibs, the aldehydes were mostly lost after drying and roasting. The main aldehydes in roasted nibs were phenyl acetaldehyde (51%, 62%) and benzaldehyde (48%, 26%) with the values in parenthesis corresponding to the amounts in the nibs fermented in the absence and presence of Natamycin, respectively. Fig. 7e and f show changes in the concentrations of these two

a 1.0E+9 8.0E+8 6.0E+8 4.0E+8 2.0E+8 0.0E+0 0

24

48

72

96

120 144





b 1.0E+9 8.0E+8 6.0E+8 4.0E+8 2.0E+8 0.0E+0 0

1.5E+8 1.0E+8 5.0E+7 0.0E+0 72

96

120

144




2.0E+8

48

e 1.5E+9 1.2E+9 9.0E+8 6.0E+8 3.0E+8 0.0E+0 24

48

72

96

120 144


d 2.0E+8 1.5E+8 1.0E+8 5.0E+7 0.0E+0 0

f 1.5E+9 1.2E+9 9.0E+8 6.0E+8 3.0E+8 0.0E+0 0

0

24

48

72

96

120 144

Fermentation time (h) 2-Heptanone_Natamycin untreated 2-Heptanone_Natamycin treated 2-Nonanone_Natamycin untreated 2-Nonanone_Natamycin treated

24 48 72 96 120 144 D

R

Fermentation time (h) Benzaldehyde_Natamycin untreated Benzaldehyde_Natamycin treated Phenylacetaldehyde_Natamycin untreated Phenylacetaldehyde_Natamycin treated



5.0E+7 4.0E+7 3.0E+7 2.0E+7 1.0E+7 0.0E+0

R


Benzaldehyde_Natamycin untreated Benzaldehyde_Natamycin treated Phenyl acetaldehyde_Natamycin untreated Phenyl acetaldehyde_Natamycin treated

g

24 48 72 96 120 144 D

Isoamyl acetate_Natamycin untreated Isoamyl acetate_Natamycin treated 2-Phenylethyl acetate_Natamycin untreated 2-Phenylethyl acetate_Natamycin treated



Isoamyl acetate_Natamycin untreated Isoamyl acetate_Natamycin treated 2-Phenylethyl acetate_Natamycin untreated 2-Phenylethyl acetate_Natamycin treated

0

R

Isoamyl alcohol_Natamycin untreated Isoamyl alcohol_Natamycin treated Phenyl ethyl alcohol_Natamycin untreated Phenyl ethyl alcohol_Natamycin treated

c

24

24 48 72 96 120 144 D


Isoamyl alcohol_Natamycin untreated Isoamyl alcohol_Natamycin treated Phenyl ethyl alcohol_Natamycin untreated Phenyl ethyl alcohol_Natamycin treated

0

81

h 5.0E+7 4.0E+7 3.0E+7 2.0E+7 1.0E+7 0.0E+0 0

24 48 72 96 120 144 D

R

Fermentation time (h) 2-Heptanone_Natamycin untreated 2-Heptanone_Natamycin treated 2-Nonanone_Natamycin untreated 2-Nonanone_Natamycin treated

Fig. 7. Changes in concentrations of main alcohols (a, b), esters (c, d), ketones (e, f) and aldehydes (g, h) in cocoa pulp (a, c, e, g) and nibs (b, d, f, h) during untreated and Natamycin-treated fermentations.

aldehydes in the pulp and nibs during fermentation. It is notable that much higher concentrations of these aldehydes were found in the nibs compared with the pulp fraction.

3.2.3.4. Ketones. Significant levels of ketones were detected in unfermented pulp and their total concentration decreased 2–5 fold during both fermentations (Fig. 6g). Four main ketones detected in fermenting

82


pulps were 2-heptanone (49%, 26%), acetophenone (28%, 43%), benzothiazole (17.5%, 17%) and 2-nonanone (5%, 13.5%). The values in brackets are their relative amounts at the end of fermentation in the absence and presence of Natamycin, respectively. Three main ketones detected in unfermented nibs were 2-heptanone, 2-nonanone and acetophenone and their total concentration increased 3 fold during fermentation in the absence and presence of Natamycin but decreased throughout drying and roasting (Fig. 6h). The relative amounts of 2heptanone, 2-nonanone and acetophenone in roasted nibs of both fermentations were in the range of 48–49%, 25–30% and 20–27%, respectively. Fig. 7g and h show the kinetics of heptanone and nonanone changes in the pulp and nibs throughout fermentation. 3.2.3.5. Pyrazines. Pyrazines were not detected in the pulp and nibs during fermentation. They were formed during roasting of the cocoa beans (Fig. 8). A total of 7 pyrazine compounds were detected with 2,3,5,6tetramethyl pyrazine, 2,5-dimethyl-3-ethylpyrazine and 2,3,5-trimethyl pyrazine being the predominant ones in both fermentations. The concentrations of these three compounds, however, were 1.5-2.5 fold higher in beans fermented in the absence of Natamycin than in those fermented in the presence of Natamycin. Roasted nibs from fermentation in the absence of Natamycin also produced significant concentration of 2-ethyl3,5-dimethyl-pyrazine which was almost 24 fold greater than in those fermented in the presence of Natamycin (Fig. 8). The profile of pyrazines found in beans fermented in March was not detected in beans fermented in October. Only 2-methyl-6-vinyl-pyrazine was detected in roasted beans from the October fermentation, and only found in beans fermented in the absence of Natamycin.

Area unit/g nib (dry basis)

3.2.4. Degradation of bean proteins and changes in the concentration of free amino acids, and alkaloids during fermentation Based on SDS–PAGE, four major protein bands of 47, 32, 28 and 18 kDa were found in unfermented nibs and changes in the intensity of these bands were monitored during fermentation (data not shown). The 47 kDa proteins were degraded during fermentation and the band representing these proteins was not observed in the gels after 72 h of fermentation while the 28 kDa proteins were degraded after 96 h. The 32 kDa proteins were partially degraded and were still detectable after 144 h and drying. The intensity of the 18 kDa protein bands was virtually unchanged throughout fermentation and drying. This pattern of protein degradation was similar for beans fermented in the absence or presence of Natamycin. Twenty one free amino acids were detected in extracts of unfermented nibs and fermented nibs (Supplementary Table 1). The total concentration of free amino acids in the unfermented nibs was

3.0E+7 2.5E+7 2.0E+7 1.5E+7 1.0E+7 5.0E+6

4.5 mg/g and this increased by 2–3 fold to approximately 11.4 mg/g in fully fermented nibs. Changes in the concentrations of total free amino acids as well as those for individual amino acids were similar for beans fermented in the absence or presence of Natamycin (no statistically significant differences p N 0.05). Greatest increases (10–16 fold) were observed for the hydrophobic amino acids (e.g., phenylalanine, leucine and tyrosine) but some non-hydrophobic acids (e.g., lysine, arginine, glycine and threonine) also exhibited such changes (data not shown). Unfermented cocoa nibs contained 15–16 mg/g of theobromine and 3–3.5 mg/g of caffeine. These concentrations remained unchanged during the first 72 h of fermentation, after which they decreased gradually to 9–10 mg/g for theobromine and 1.5–1.6 mg/g for caffeine at the end of fermentation. The total loss of theobromine and caffeine was in the range of 38–40% and 50–54%, respectively, throughout the fermentation process. These changes were the same for fermentations in the absence or presence of Natamycin. (+)-Catechin and (−)-epicatechin were not detected (b 0.04 mg/g) in either unfermented nibs or fermented nibs. 3.3. Quality evaluation of cocoa beans During fermentation, the external appearance of the cocoa pulp changed differently for the two fermentations. The mucilage of the beans fermented in the absence of Natamycin appeared dry and only slightly sticky while that of beans fermented in the presence of Natamycin was more moist and sticky and less degraded. The dried beans from both fermentations had similar moisture contents at approximately 7–7.5%. Dried beans from fermentations where yeasts were inhibited were heavier and contained significantly greater proportion (3–7%) of shell material than beans from the control fermentations (p b 0.05). For three trials, August 2010, March 2011 and October 2011, the percentage shell for beans fermented in the absence and presence of Natamycin was, respectively, 11.79% and 20.24%, 14.29% and 17.53% and 20.30% and 28.96%. As evaluated by the cut test, all 30 beans examined from fermentations in the absence of Natamycin exhibited a full brown color that was indicative of successful fermentation. However, beans fermented in the presence of Natamycin retained a purplish color, suggesting unsuccessful fermentation. Consequently, chocolate made from beans where yeasts did not grow (Natamycin present) had a lighter brown color. Statistical analysis (t-test) of the mean liking scores revealed that chocolate made from beans fermented in the absence of Natamycin (yeast growth) was liked significantly (p b 0.05) more than chocolate prepared from beans fermented in the presence of Natamycin (no yeast growth). Furthermore, the latter tasted more acidic and lacked characteristic chocolate flavor, whereas the former gave typical chocolate characters. These conclusions were found for three independent fermentation trials using beans harvested on three different occasions (Fig. 9). The beans harvested in October 2011 gave an inferior tasting chocolate (Fig. 9c) compared to the other two batches (Fig. 9a, b), but the differences between beans fermented with or without yeasts were still significant. 4. Discussion

0.0E+0

Fig. 8. The concentration of pyrazines produced during roasting of cocoa nibs fermented in the absence ( ) and presence of Natamycin ( ).

Natamycin is an inhibitor of yeast viability and growth and has been approved for use at concentrations up to 2000 ppm to control yeast spoilage in some foods (Delves-Broughton et al., 2005). In this study, Natamycin was used to inhibit yeast growth in cocoa bean fermentations, thereby providing a mechanism to investigate the role and contribution of yeasts to this fermentation and their impact on chocolate quality. When added to freshly extracted cocoa beans at concentrations of 100 ppm or more, it inactivated indigenous or contaminating yeasts and prevented their growth during subsequent fermentation. It did not affect the growth of bacteria during these fermentations where


83

Fig. 9. Sensory evaluation of chocolates made from cocoa beans fermented in the absence of Natamycin ( ), and in the presence of Natamycin ( ); beans harvested in August 2010 (a), March 2011 (b) and October 2011 (c).

their population levels and species profiles were similar to those found in fermentations with no added Natamycin.

4.1. Microbial ecology of fermentations Control cocoa bean fermentations, where no Natamycin was added, gave successional growth of indigenous yeast and bacterial species similar to that reported in many previous studies (Lima et al., 2011; Schwan and Wheals, 2004; Thompson et al., 2013). In the case of yeasts, H. guilliermondii dominated the first 24–48 h of fermentation and this was followed by the growth and prevalence of P. kudriavzevii and then K. marxianus. We did not observe the presence of S. cerevisiae as found by Dircks (2009) for fermentation of North Queensland cocoa beans or beans fermented in Brazil (Pereira et al., 2012, 2013), Indonesia (Ardhana and Fleet, 2003) and sometimes, in Ghana (Daniel et al., 2009; Jespersen et al., 2005). The reasons for such discrepancies, which have been observed by others (Lima et al., 2011), could be many and require systematic investigation. However, none of the yeast species just mentioned or any other yeast species were found for cocoa beans fermented in the presence of 100–200 ppm of Natamycin. L. plantarum was the main species found in our cocoa fermentations but L. fermentum and Lactococcus lactis were also observed (October fermentation), as reported elsewhere (Camu et al., 2008b; Illeghems et al., 2012; Kostinek et al., 2008; Nielsen et al., 2007; Papalexandratou et al., 2011a). The population dynamics for lactic acid bacteria were similar for fermentations in the presence or absence of yeasts. The acetic acid bacteria associated with the fermentations were A. pasteurianus and G. frateurii. While A. pasteurianus is the main species often found in cocoa fermentations (Ardhana and Fleet, 2003; Camu et al., 2008b; Carr et al., 1979; Nielsen et al., 2007; Papalexandratou et al., 2011b), G. oxydans is also frequently detected (Camu et al., 2008b; Nielsen et al., 2007). G. oxydans was not found in our study and our isolates of G. frateurii gave RFLP patterns and 16S rDNA sequences distinct from a reference strain of G. oxydans. The beans from the October fermentation did not have a strong presence of acetic acid bacteria and the reasons for this are not evident. It is notable that chocolate prepared from these beans was less-liked in sensory evaluation (Fig. 9). It is generally accepted that the growth of acetic acid bacteria in cocoa fermentations is linked to the growth of yeasts and oxidation of ethanol produced by such growth (Schwan et al., 1995; Schwan and

Wheals, 2004; Thompson et al., 2013). However, this view needs reconsideration because these bacteria grew in fermentations with added Natamycin, where no yeasts developed and little ethanol was produced. Their growth in these fermentations was similar to that of the control ferments. Depending on fermentation hygiene and length, Bacillus species sometimes grow during the later stages of cocoa fermentation (Lima et al., 2011; Schwan and Wheals, 2004; Thompson et al., 2013). While no Bacillus species were detected in beans of the March fermentation, low populations of Bacillus subtilis were found at the end of the October fermentation. Their presence was similar for fermentations in the presence or absence of Natamycin.

4.2. Changes in chemical components during fermentation 4.2.1. Sugars Pulp sugars provide the substrates that drive the microbial fermentation and create the environmental conditions that stimulate the production of chocolate flavor precursors within the bean. Fructose (approx. 70 mg/g) and glucose (approx. 50 mg/g) were the main fermentable sugars in freshly extracted pulp. These concentrations are consistent with those reported in early studies (Roelofsen, 1958) and more recent works (Galvez et al., 2007; Nielsen et al., 2007; Papalexandratou et al., 2011b). In accordance with these studies, fructose and glucose were largely utilized throughout control fermentations where yeasts were active. In contrast, these sugars were less utilized when yeast growth was inhibited, leaving higher residuals of fructose (16.5 mg/g) and glucose (35.5 mg/g) in the pulp environment with potential for diffusion into the nibs. The concentration of reducing sugars in fermented nibs is important because they are involved in the development of chocolate flavor through Maillard reactions with amino acids during roasting (Afoakwa et al., 2008; Redgwell et al., 2003). Unfermented cocoa nibs contained sucrose (17 mg/g) as the main sugar and low concentrations of glucose (0.5 mg/g) and fructose (1 mg/g). Nib sucrose was completely hydrolyzed during fermentations in the absence and presence of Natamycin, leading to progressive increases in the concentrations of fructose and glucose. Such trends are well reported in the literature (Hashim et al., 1998; Pereira et al., 2012; Reineccius et al., 1972; Rohan and Stewart, 1967a) and are attributed to invertase activity within the nib (Hansen

84


et al., 1998). The final concentrations of glucose and fructose in the nib were less than those expected of complete hydrolysis of the initial amount of sucrose, an observation also reported by Reineccius et al. (1972) and Hansen et al. (1998). This finding suggests that some of these sugars may diffuse out of the nibs during fermentation or may be further metabolized into other products. Although the residuals of fructose and glucose in the pulp fractions were significantly higher in the yeast inhibited fermentation, our data suggest that these sugars do not diffuse into the nibs which was a conclusion also noted by Hansen et al. (1998). This observation is noteworthy because the concentrations of reducing sugars in the nib have a major impact on the development of chocolate flavor (Reineccius et al., 1972; Rohan and Stewart, 1967a). 4.2.2. Ethanol, glycerol and mannitol Ethanol is a major metabolite of pulp fermentation and this has been recognized since early studies when it was considered to be a primary agent of bean death and initiation of reactions leading to the development of chocolate flavor (Roelofsen, 1958). Its production correlates with the growth of yeasts and their fermentation of pulp sugars. This relationship was clearly demonstrated in our studies where inhibition of yeast growth gave little ethanol presence in the bean pulp or nibs. In control fermentation where yeasts grew, ethanol concentration in the pulp increased and then decreased as fermentation progressed. These kinetics were paralleled by its movement into and out of the bean nib. Such behavior has been reported previously (Camu et al., 2008a; Galvez et al., 2007; Lefeber et al., 2012; Papalexandratou et al., 2011a; Roelofsen, 1958). Its decline in the later stages of fermentation is thought to be due to its oxidation to acetic acid by acetic acid bacteria, and loss due to evaporation as the temperature of the cocoa mass increases (Galvez et al., 2007; Schwan and Wheals, 2004). The maximum concentrations of ethanol reported in the pulp and nib fractions varies from as little as 5–20 mg/g as found in our study and elsewhere (Camu et al., 2008b; Galvez et al., 2007; Lefeber et al., 2012; Papalexandratou et al., 2011a) to values as high as 30–60 mg/g (Ardhana and Fleet, 2003; Roelofsen, 1958; Schwan, 1998). Such discrepancies might be explained by variations in the initial amounts of sugars present in the pulp, the species of yeasts that develop during fermentation and lack of standardization in the methods used to measure and report its presence in cocoa pulp and nibs. Fermentations that develop a stronger presence of highly fermentative yeasts such as S. cerevisiae are more likely to have a greater concentration of ethanol (Ardhana and Fleet, 2003; Papalexandratou et al., 2011a; Schwan, 1998) than those which develop less-fermentative yeasts, such as species of Hanseniaspora, Pichia and Kluyveromyces (Galvez et al., 2007). S. cerevisiae was not found in our control cocoa fermentations and this may account for ethanol concentrations at the lower end of the range. According to Roelofsen and Giesberger (1947) and Quesnel (1965), ethanol concentrations in excess of 40 mg/g are needed to kill the bean. Such concentrations are not achieved in many fermentations and it is now considered that acetic acid is the primary agent causing bean death, although its effect is ameliorated by the production of ethanol and increasing temperature (Quesnel, 1965). Glycerol is an important secondary product of yeast metabolism of sugars and contributes to sensory properties of mouth-feel and sweet taste (Swiegers et al., 2005). Surprisingly, glycerol production during cocoa bean fermentations has not been reported in the past. We observed its production in the pulp but it did not transfer into the nib, so its formation during fermentation was unlikely to impact on chocolate quality. The formation of glycerol was due to yeast growth since it was not produced in fermentations containing Natamycin. In recent years, several authors have reported the production of mannitol in the pulp (9–35 mg/g) and nibs (1.5-5 mg/g) during cocoa bean fermentation (Camu et al., 2008b; Papalexandratou et al., 2011a, b). This is linked to the growth of L. fermentum and its ability to metabolize fructose directly to mannitol (De Vuyst et al., 2010; Lefeber et al., 2010; Saha and Racine, 2011). We confirmed this behavior during

fermentation of the October beans where L. fermentum was present. Its production was not affected by the presence of Natamycin. Mannitol has a sweet and cool taste (Soetaert et al., 1999) so its presence in the nibs could possibly contribute to sensory characteristics of chocolate. 4.2.3. Organic acids Citric, acetic and lactic acids have key roles in the development of chocolate character (De Vuyst et al., 2010; Holm et al., 1993; Roelofsen, 1958). Apart from their direct impacts on sensory quality, their movement in and out of the nib modulates bean pH which indirectly affects the activity of bean enzymes such as proteases and invertase involved in the production of free amino acids, peptides and reducing sugars, all of which impact on chocolate flavor (De Vuyst et al., 2010; Hansen et al., 1998; Thompson et al., 2013). Citric acid is the main acid in unfermented cocoa pulp (5–40 mg/g) and nibs (2–9 mg/g) (Ardhana and Fleet, 2003; Camu et al., 2008a; Galvez et al., 2007; Lefeber et al., 2011; Papalexandratou et al., 2011b). The kinetics of its utilization in the pulp or nibs were similar for fermentations in the presence or absence of yeasts. Although there are general reports that yeasts utilize pulp citric acid (Schwan and Wheals, 2004), the evidence for this was not found in our study. Its utilization is more connected with the growth of certain species of lactic acid bacteria such as Lactococcus lactis and L. fermentum (Fig. 2d, f), but not L. plantarum as reported also by Lefeber et al. (2010). However, changes in its concentration in the pulp did not manifest as changes to its concentration in the nibs, suggesting an inability of this acid to freely move into the nibs, as also suggested by De Vuyst et al. (2010). Consequently, it probably undergoes independent metabolism in the nibs. The production of acetic acid at levels of 1–2% during fermentation is necessary to kill the bean (Quesnel, 1965) and it is generally considered that this production mainly comes from the oxidation of ethanol by acetic acid bacteria (De Vuyst et al., 2010; Schwan and Wheals, 2004; Lima et al., 2011; Thompson et al., 2013). Fermentation in the absence of yeasts did not produce ethanol, but acetic acid at maximum concentrations of 13.7 mg/g in the pulp and 15.9 mg/g in the nibs were found and were similar to concentrations obtained when yeasts contributed to the fermentation (Fig. 5). It may be concluded, therefore, that this acid may also come from other metabolic mechanisms, not involving the oxidation of yeast ethanol. Roelofsen (1958) and Roelofsen and Giesberger (1947) mentioned in early studies that lactic acid bacteria could be a main source of acetic acid in cocoa fermentations and this now seems to be the case. Both hetero and homofermentative lactic acid bacteria can generate acetic acid from fermentable sugars under particular conditions as well as from the transformation of fructose to mannitol and the metabolism of citrate (De Vuyst et al., 2010; Palles et al., 1998; Smetankova et al., 2012; Zalan et al., 2010). The higher concentrations of acetic acid found in the October fermentations is likely to come from the greater contribution of lactic acid bacteria (Fig. 2d, f), especially L. fermentum (Annan et al., 2003). It is also relevant to point out two other possible mechanisms for acetic acid production. First, apart from oxidizing ethanol, acetic acid bacteria may also generate this acid by direct metabolism of sugars (Awad et al., 2012). Second, yeasts are also well known to produce acetic acid by sugar metabolism (Swiegers et al., 2005). Ethanol production by yeasts probably speeds up the rate of acetic acid formation and consequently bean death, but ethanol production may not be necessary to achieve the acetic acid concentrations required for bean death. Fermentation of beans in the absence of yeasts always gave pulps and nibs with a final pH less than that when yeasts were participants (Fig. 3). It was reasonable to expect that, in the absence of yeast growth, there would be a stronger fermentation of pulp sugars by lactic acid bacteria and production of lactic acid. This did not occur (Fig. 5) and changes in the concentrations of lactic acid in the pulp and nibs were similar for both control and Natamycin treated fermentations. The kinetics of lactic acid production and concentrations in the pulp (25–29 mg/g) and accumulation in the nibs (10–13 mg/g) were similar to those


reported elsewhere (Camu et al., 2007; Papalexandratou et al., 2011b; Pereira et al., 2012). It is assumed that lactic acid in the nibs originates by diffusion from the pulp environment as suggested by Lehrian and Paterson (1983) and De Vuyst et al. (2010) and not by other metabolic mechanisms that might occur within the nib. 4.2.4. Yeasts and quality of bean and chocolate The chemistry of chocolate flavor is complex, with over 500 nonvolatile and volatile components contributing to its character (Afoakwa et al., 2008; Bonvehi, 2005; Counet et al., 2002; Frauendorfer and Schieberle, 2006). Some of these components will be intrinsic to the bean, some will be generated by bean metabolism after its death, some will originate as metabolites of microbial species associated with fermentation and some, probably most, will be generated by the Maillard and other reactions that occur during bean roasting. Here, we focus on the role of yeasts in contributing to these mechanisms, directly or indirectly. Directly, yeast fermentation of pulp sugars, produces a vast array of volatile metabolites (e.g., higher alcohols, organic acids, esters, aldehydes, ketones, etc.) that are well known for their aromatic and flavorant properties (Swiegers et al., 2005). In the case of cocoa bean fermentation, it is not clear how such volatiles might impact on chocolate flavor because, firstly, they must diffuse into the beans and, secondly, it is expected that they would be mostly lost by evaporation or otherwise transformed during the roasting operation. Although no experimental evidence has been given, it is often mentioned in the literature that such metabolites might contribute unique estery, fruity, floral and other notes to chocolate character (Aculey et al., 2010; Afoakwa et al., 2008; Lima et al., 2011; Roelofsen, 1958). For this reason, we studied the evolution of volatile metabolites in the cocoa pulp and their transfer to the bean nibs during fermentation. We are not aware of any other studies on these kinetics, although Rodriguez-Campos et al. (2011, 2012) recently reported the volatile levels in bean nibs during the course of fermentation. All of the higher alcohols, esters, aldehydes and ketones described in our results section have been mentioned in lists of key volatiles that contribute to chocolate aroma and flavor, the most notable found in our study being phenylacetaldehyde, benzaldehyde, phenylethanol, 3-methyl-1-butanol, phenylethylacetate and 2heptanone (Bonvehi, 2005; Counet et al., 2002; Frauendorfer and Schieberle, 2006; Jinap et al., 1998). Absence of the esters, ethyl acetate and isoamyl acetate (also known as isopentyl acetate) from these lists is notable, although they were produced throughout our fermentations (Fig. 7c, d). The production of these esters as well as the other volatiles was also reported during the fermentations conducted by Aculey et al. (2010) and Rodriguez-Campos et al. (2011, 2012). Almost no higher alcohols or esters were found in pulps or beans fermented in the absence of yeasts (Fig. 6a, b), thereby demonstrating that yeasts and not the lactic acid bacteria or acetic acid bacteria are responsible for their production and any impact on chocolate flavor. With respect to aldehydes, there was little production in the pulp, compared to the bean, leading to the conclusion that these volatiles are mostly generated during the course of fermentation by biochemical reactions within the bean. This conclusion is supported by the observation that, for the most part, aldehyde contents in the bean nibs were similar for fermentations conducted in the presence or absence of yeasts (Fig. 6e, f). Conclusions about the ketones were similar but complicated by the fact that both the pulp and beans had high initial levels before fermentation and changes during fermentation, at least for the two prominent ones (2-heptanone and 2-nonanone), were not affected by microbial action. Pyrazines are major volatiles that contribute to chocolate flavor (Bonvehi, 2005; Frauendorfer and Schieberle, 2006). Accordingly, pyrazines were found in our fermented beans after roasting (Fig. 8) and the profiles obtained (dominance of 2,3,5,6, tetramethyl pyrazine) are consistent with other reports (Afoakwa et al., 2008; Bonvehi, 2005; Frauendorfer and Schieberle, 2006; Reineccius et al., 1972).

85

Significantly less pyrazines were found in beans fermented in the absence of yeasts and these beans gave chocolate that was judged to be weak in chocolate characters. Pyrazines are predominantly formed during roasting by reactions between reducing sugars and free amino acids (Afoakwa et al., 2008; Hashim et al., 1998; Jinap et al., 1998; Reineccius et al., 1972; Rohan and Stewart, 1967a). Consequently, the role of yeasts in contributing to the levels of reducing sugars and free amino acids within the bean nibs will be important. The presence of reducing sugars has already been discussed. The concentration of free amino acids within the beans increases during fermentation and is attributed to the activity of endo and exo proteases that degrade bean storage proteins (Biehl et al., 1985; Hansen et al., 1998; Kirchhoff et al., 1989; Rohan and Stewart, 1967b). This behavior was confirmed in our study where key storage proteins were degraded, as observed previously (Amin et al., 1997; Buyukpamukcu et al., 2001; Zak and Keeney, 1976) and the concentration of free amino acids increased by 2–3 fold throughout fermentation. The magnitude of this change, the profile of amino acids and substantial increases in the concentrations of hydrophobic amino acids (e.g., phenylalanine, leucine and tyrosine) that we found were similar to data reported by others (Hashim et al., 1998; Kirchhoff et al., 1989; Maravalhas, 1972; Rohsius et al., 2006). However, yeasts did not specifically contribute to these changes in proteins and amino acids because they were similar for fermentations in the presence or absence of yeasts. Significantly, roasted beans from fermentations in the absence of yeasts gave less pyrazines and this correlated with sensory data that these beans had less chocolate character (Fig. 9). At this stage, we are unable to explain this difference on the basis of availability of reducing sugars and free amino acids for pyrazine formation. Pyrazines are not present in unfermented pulp or beans but several studies report some production during fermentation and have correlated this with the growth of Bacillus species (Hashim et al., 1998; Zak and Keeney, 1976). There are no studies that link their de-novo production by yeasts or other bacteria during cocoa fermentation. We did not detect any production of pyrazines during our fermentations, either in the presence or absence of yeasts, and we may conclude that the yeasts, lactic acid bacteria and acetic acid bacteria associated with these fermentations do not synthesize pyrazines. Bacillus spp. were not detectable in our March fermentations and only a very weak presence of Bacillus subtilis (102-103 cfu/g) was detected in the later stages of the October fermentations (Fig. 2d, f). In addition to production of free amino acids, the breakdown of bean proteins generates various peptides that are considered to contribute specific chocolate flavors (Biehl et al., 1985; Voigt et al., 1994a,b. The formation of these peptides arises from the action of particular proteases, the combined activity of which is considered to produce the strongest chocolate flavor in the pH range 5.0–5.5 (Biehl et al., 1985). In agreement with previous studies (Ardhana and Fleet, 2003; Guehi et al., 2010; Nazaruddin et al., 2006), nib pH decreased during fermentation from initial values of 6.5–7.0 to values of 4.5–5.5 (Fig. 3). Bean nib pH was always slightly lower when yeast growth during fermentation was inhibited, but the differences were small and probably not sufficient to enable detection of quantitative differences in protein and amino acid changes within the nib. This observation was also confirmed by following the changes in bean proteins by electrophoretic analyses. Weaker chocolate flavor is reported when proteolysis occurs at pH values below 5.0 (Biehl et al., 1985) and this may explain the weaker chocolate character observed for the product derived from the October fermented beans. The polyphenols, epicatechin (10–40 mg/g) and catechin (0.2– 10 mg/g) and the alkaloids, theobromine (1–25 mg/g) and caffeine (0.06–8 mg/g) are important components of unfermented cocoa beans that contribute bitterness and astringency to chocolate character (Camu et al., 2008a; Kim and Keeney, 1984; Lefeber et al., 2012; Nazaruddin et al., 2006; Payne et al., 2010). Their concentrations decrease by about 10–20% during fermentation due to diffusion out of the nibs (Camu et al., 2008a; Kim and Keeney, 1984). Our findings

86


confirmed this behavior for theobromine and caffeine and demonstrated no impact of yeasts on these components. We could not draw conclusions about the impact of yeasts on the polyphenols, epicatechin and catechin, because only trace amounts (b 0.04 mg/g) were detected in our beans. Such data are inconsistent with epicatechin and catechin levels that have been reported in the literature as previously referenced. We confirmed these low levels of polyphenols using an additional extraction procedure (Camu et al., 2008a) and verified the validity of our analytical procedures with samples spiked with known amounts of epicatechin. The cultivar of our cocoa beans originated from Papua New Guinea where very low, but measurable, levels of epicatechin and catechin have been reported by other researchers (Counet et al., 2004). Based on visual and other sensory criteria, cocoa beans fermented in the absence of yeasts were clearly different. They had a higher content of shell material suggesting that the pulp had not been fully degraded and remained attached to the testa. For acceptable dried beans, shell content should not exceed 12–16% of the bean dry weight (Fowler, 2009). Beans fermented in the absence of yeasts did not meet this criterion whereas control fermented beans, where yeasts were active, met this specification. These observations confirm an important role of yeasts, such as K. marxianus, as found in our studies, in degrading pulp pectins (Leal et al., 2008). The nibs of beans fermented in the absence of yeasts were purplish-violet in color and not fully brown, and would not pass the cut test with respect to this property (Dand, 1996). Nibs of this color are considered under-fermented. It could be concluded that the bean glycosidase activity necessary to convert the bean anthocyanins (purple–red) to the colorless form (Pettipher, 1986; Wollgast and Anklam, 2000) and the bean polyphenoloxidase activity needed to transform phenolic compounds to brown polymers (Kyi et al., 2005) may not have their full functionality for fermentation in the absence of yeasts. We can only speculate that the presence of ethanol may be needed to facilitate the activity of these enzymes, as reported for the action of some other glycosidases (Grimaldi et al., 2005). Limited oxygen availability due to incomplete degradation of the pulp layer may be another factor affecting the action of these enzymes. Further research is needed to investigate these possibilities. Chocolate prepared from beans fermented without yeasts had a lighter brown appearance, had less intense chocolate flavor and was less preferred than that prepared from beans fermented in the presence of yeasts.

5. Conclusion Based on shell content, nib color and chocolate sensory criteria, we conclude that yeast growth and activity are essential for successful cocoa bean fermentation. From an analytical perspective, the main differences when yeast growth was inhibited were: increased shell content, the absence of ethanol, higher alcohol and ester production throughout fermentation and lesser presence of pyrazines in the roasted product. Several yeast species (e.g., H. guilliermondii, P. kudriavzevii, S. cerevisiae, K. marxianus) are consistently found to be active in cocoa bean fermentations throughout the world. Further research is needed to determine if the action of all these species is essential for a successful fermentation and how individual species affect the process and final chocolate quality. Such information will then guide the selection and development of defined starter cultures for a better controlled process. Some studies with inoculated strains of H. guilliermondii and P. kudriavzevii (Dircks, 2009) and Pichia kluyveri and K. marxianus (Crafack et al., 2013) have already demonstrated specific influences on chocolate flavor. Our findings also demonstrate the need for more research to better understand the biochemical reactions that occur within the bean during fermentation, especially those that affect pH and acidity such as citric acid degradation, the production of reducing sugars and amino acids that provide the precursors to pyrazine production, and the transformation of polyphenols. The impact of ethanol, as produced by yeasts, on these reactions requires investigation.

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijfoodmicro.2013.12.014. Acknowledgements The authors gratefully acknowledge Mr. Yan Diczbalis, Department of Agriculture and Forestry, Queensland, Australia and Mr. Alan Mortimer, the Australian Blending Company for assisting in supplying cocoa beans for this project. Ms. Van Ho thanks the Department of Education, Employment and Workplace Relations, Australian Government for providing an Endeavour PhD Scholarship to conduct this research. References Aculey, P.C., Snitkjaer, P., Owusu, M., Bassompiere, M., Takrama, J., Norgaard, L., Petersen, M.A., Nielsen, D.S., 2010. Ghanaian cocoa bean fermentation characterized by spectroscopic and chromatographic methods and chemometrics. J. Food Sci. 75, S300–S307. Afoakwa, E.O., Paterson, A., Fowler, M., Ryan, A., 2008. Flavor formation and character in cocoa and chocolate: a critical review. Crit. Rev. Food Sci. Nutr. 48, 840–857. Amin, I., Jinap, S., Jamilah, B., 1997. Vicilin-class globulins and their degradation during cocoa fermentation. Food Chem. 59, 1–5. Annan, N.T., Poll, L., Sefa-Dedeh, S., Plahar, W.A., Jakobsen, M., 2003. Volatile compounds produced by Lactobacillus fermentum, Saccharomyces cerevisiae and Candida krusei in single starter culture fermentations of Ghanaian maize dough. J. Appl. Microbiol. 94, 462–474. Ardhana, M.M., Fleet, G.H., 2003. The microbial ecology of cocoa bean fermentations in Indonesia. Int. J. Food Microbiol. 86, 87–99. Awad, H.S., Diaz, R., Malek, R.A., Othman, N., Aziz, R.A., El Enshasy, H.A., 2012. Efficient production process for food grade acetic acid by Acetobacter aceti in shake flask and in bioreactor cultures. Egypt. J. Chem. 9, 2275–2286. Biehl, B., Brunner, E., Passern, D., Quesnel, V.C., Adomako, D., 1985. Acidification, proteolysis and flavor potential in fermenting cocoa beans. J. Sci. Food Agric. 36, 583–598. Bonvehi, J., 2005. Investigation of aromatic compounds in roasted cocoa powder. Eur. Food Res. Technol. 221, 19–29. Buyukpamukcu, E., Goodall, D.M., Hansen, C.E., Keely, B.J., Kochlar, S., Wille, H., 2001. Characterization of peptides formed during fermentation of cocoa bean. J. Agric. Food Chem. 49, 5822–5827. Camu, N., De Winter, T., Verbrugghe, K., Cleenwerck, I., Vandamme, P., Takrama, J.S., Vancanneyt, M., De Vuyst, L., 2007. 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. 73, 1809–1824. Camu, N., De Winter, T., Addo, S.K., Takrama, J.S., Bernaert, H., De Vuyst, L., 2008a. Fermentation of cocoa beans: influence of microbial activities and polyphenol concentrations on the flavor of chocolate. J. Sci. Food Agric. 88, 2288–2297. Camu, N., Gonzalez, A., De Winter, T., Van Schoor, A., De Bruyne, K., Vandamme, P., Takrama, J.S., Addo, S.K., De Vuyst, L., 2008b. Influence of turning and environmental contamination on the dynamics of populations of lactic acid and acetic acid bacteria involved in spontaneous cocoa bean heap fermentation in Ghana. Appl. Environ. Microbiol. 74, 86–98. Carpenter, R.P., Lyon, D.H., Hasdell, T.A., 2000. Guidelines for sensory analysis in food product development and quality control, 2nd ed. An Aspen Publication. Carr, J.G., Davies, P.A., Dougan, J., 1979. Cocoa fermentation. J. Sci. Food Agric. 30, 841-841. Cocolin, L., Rantsiou, K., Iacumin, L., Cantoni, C., Comi, G., 2002. Direct identification in food samples of Listeria spp. and Listeria monocytogenes by molecular methods. Appl. Environ. Microbiol. 68, 6273–6282. Cooper, K.A., Campos-Gimenez, E., Alvarez, D.J., Nagy, K., Donovan, J.L., Williamson, G., 2007. Rapid reversed phase ultra-performance liquid chromatography analysis of the major cocoa polyphenols and inter-relationship of their concentrations in chocolate. J. Agric. Food Chem. 55, 2841–2847. Counet, C., Callemien, D., Ouwerx, C., Collin, S., 2002. Use of gas chromatographyolfactometry to identify key odorant compounds in dark chocolate. Comparison of samples before and after conching. J. Agric. Food Chem. 50, 2385–2391. Counet, C., Ouwerx, C., Rosoux, D., Collin, S., 2004. Relationship between procyanidin and flavor contents of cocoa liquors from different origins. J. Agric. Food Chem. 52, 6243–6249. Crafack, M., Mikkelsen, M.B., Saerens, S., Knudsen, M., Blennow, A., Lowor, S., Takrama, J., Swiegers, J.H., Petersen, G.B., Heimdal, H., Nielsen, D.S., 2013. Influencing cocoa flavor using Pichia Kluyveri and Kluyveromyces marxianus in a defined mixed starter culture for cocoa fermentation. Int. J. Food Microbiol. 167, 103–116. Dand, R., 1996. The international cocoa trade, 2nd ed. Woodhead Publishing Limited, New York. Daniel, H.M., Vrancken, G., Takrama, J.F., Camu, N., De Vos, P., De Vuyst, L., 2009. Yeast diversity of Ghanaian cocoa bean heap fermentations. FEMS Yeast Res. 9, 774–783. De Vuyst, L., Lefeber, T., Papalexandratou, Z., Camu, N., 2010. The functional role of lactic acid bacteria in cocoa bean fermentation. In: Mozzi, F., Raya, R.R., Vignolo, G.M. (Eds.), Biotechnology of Lactic Acid Bacteria — Novel Applications. Wiley-Blackwell, USA, pp. 301–325. Delves-Broughton, J., Thomas, L.V., Doan, C.H., Davidson, P.M., 2005. Natamycin, In: Davidson, P.M., Sofos, J.N., Branen, A.L. (Eds.), Antimicrobials in foods, 3rd ed. CRC Taylor & Francis, USA, pp. 275–289.

V.T.T. Ho et al. / International Journal of Food Microbiology 174 (2014) 72–87 Dircks, H.D., 2009. Investigation into the fermentation of Australian cocoa beans and its effect on microbiology, chemistry and flavor. The University of New South Wales, Australia. Esteve-Zarzoso, B., Belloch, C., Uruburu, F., Querol, A., 1999. Identification of yeasts by RFLP analysis of the 5.8S rRNA gene and the two ribosomal internal transcribed spacers. Int. J. Syst. Bacteriol. 49, 329–337. European Food Safety Authority, 2009. Scientific opinion on the use of natamycin (E 235) as a food additive. EFSA J. 7, 1412. Fowler, M.S., 2009. Cocoa beans: from tree to factory, In: Beckett, S.T. (Ed.), Industrial chocolate manufacture and use, 4th ed. Blackwell Science, Oxford, pp. 10–47. Frauendorfer, F., Schieberle, P., 2006. Identification of the key aroma compounds in cocoa powder based on molecular sensory correlations. J. Agric. Food Chem. 54, 5521–5529. Galvez, S.L., Loiseau, G., Paredes, J.L., Barel, M., Guiraud, J.P., 2007. Study on the microflora and biochemistry of cocoa fermentation in the Dominican Republic. Int. J. Food Microbiol. 114, 124–130. Granchi, L., Bosco, M., Messini, A., Vincenzini, M., 1999. Rapid detection and quantification of yeast species during spontaneous wine fermentation by PCR-RFLP analysis of the rDNA ITS region. J. Appl. Microbiol. 87, 949–956. Grimaldi, A., Bartowsky, E., Jiranek, V., 2005. A survey of glycosidase activities of commercial wine strains of Oenococcus oeni. Int. J. Food Microbiol. 105, 233–244. Guehi, S.T., Dabonne, S., Ban-Koffi, L., Kedjebo, K., Zahouli, G.E.B., 2010. Effect of turning beans and fermentation method on the acidity and physical quality of raw cocoa beans. Adv. J. Food Sci. Technol. 2, 163–171. Hansen, C.E., del Olmo, M., Burri, C., 1998. Enzyme activities in cocoa beans during fermentation. J. Sci. Food Agric. 77, 273–281. Hashim, P., Selamat, J., Muhammad, S.K.S., Ali, A., 1998. Effect of mass and turning time on free amino acid, peptide-N, sugar and pyrazine concentration during cocoa fermentation. J. Sci. Food Agric. 78, 543–550. Holm, C.S., Aston, J.W., Douglas, K., 1993. The effects of the organic-acids in cocoa on the flavor of chocolate. J. Sci. Food Agric. 61, 65–71. Illeghems, K., De Vuyst, L., Papalexandratou, Z., Weckx, S., 2012. Phylogenetic analysis of a spontaneous cocoa bean fermentation metagenome reveals new insights into its bacterial and fungal community diversity. PLoS ONE 7. Jespersen, L., Nielsen, D.S., Honholt, S., Jakobsen, M., 2005. Occurrence and diversity of yeasts involved in fermentation of West African cocoa beans. FEMS Yeast Res. 5, 441–453. Jinap, S., Rosli, W.I.W., Russly, A.R., Nordin, L.M., 1998. Effect of roasting time and temperature on volatile component profiles during nib roasting of cocoa beans (Theobroma cacao). J. Sci. Food Agric. 77, 441–448. Kim, H., Keeney, P.G., 1984. (−)-Epicatechin content in fermented and unfermented cocoa beans. J. Food Sci. 49, 1090–1092. Kirchhoff, P.M., Biehl, B., Crone, G., 1989. Peculiarity of the accumulation of free aminoacids during cocoa fermentation. Food Chem. 31, 295–311. Kostinek, M., Ban-Koffi, L., Ottah-Atikpo, M., Teniola, D., Schillinger, U., Holzapfel, W.H., Franz, C., 2008. Diversity of predominant lactic acid bacteria associated with cocoa fermentation in Nigeria. Curr. Microbiol. 56, 306–314. Kyi, T.M., Daud, W.R.W., Mohammad, A.B., Wahid Samsudin, M., Kadhum, A.A.H., Talib, M.Z.M., 2005. The kinetics of polyphenol degradation during the drying of Malaysian cocoa beans. Int. J. Food Sci. Technol. 40, 323–331. Leal, G.A., Gomes, L.H., Efraim, P., Tavares, F.C.A., Figueira, A., 2008. Fermentation of cacao (Theobroma cacao L.) seeds with a hybrid Kluyveromyces marxianus strain improved product quality attributes. FEMS Yeast Res. 8, 788–798. Lefeber, T., Janssens, M., Camu, N., De Vuyst, L., 2010. 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. 76, 7708–7716. Lefeber, T., Gobert, W., Vrancken, G., Camu, N., De Vuyst, L., 2011. Dynamics and species diversity of communities of lactic acid bacteria and acetic acid bacteria during spontaneous cocoa bean fermentation in vessels. Food Microbiol. 28, 457–464. Lefeber, T., Papalexandratou, Z., Gobert, W., Camu, N., De Vuyst, L., 2012. On-farm implementation of a starter culture for improved cocoa bean fermentation and its influence on the flavour of chocolates produced thereof. Food Microbiol. 30, 379–392. Lehrian, D.W., Paterson, G.R., 1983. Cocoa fermentation. In: Reed, G. (Ed.), Biotechnology, a comprehensive treatise, vol. 5. Verlag Chemie, Switzerland, pp. 529–575. Lima, L.J.R., Almeida, M.H., Nout, M.J.R., Zwietering, M.H., 2011. Theobroma cacao L., “The Food of the Gods”: quality determinants of commercial cocoa beans, with particular reference to the impact of fermentation. Crit. Rev. Food Sci. Nutr. 51, 731–761. Maravalhas, N., 1972. Amino acids in fermented and unfermented cocoa beans. Rev. Int. Chocolaterie 27, 22–23. Nazaruddin, R., Seng, L.K., Hassan, O., Said, M., 2006. Effect of pulp preconditioning on the content of polyphenols in cocoa beans (Theobroma Cacao) during fermentation. Ind. Crop. Prod. 24, 87–94. Nielsen, D.S., Honholt, S., Tano-Debrah, K., Jespersen, L., 2005. Yeast populations associated with Ghanaian cocoa fermentations analyzed using denaturing gradient gel electrophoresis (DGGE). Yeast 22, 271–284. Nielsen, D.S., Teniola, O.D., Ban-Koffi, L., Owusu, M., Andersson, T.S., Holzapfel, W.H., 2007. The microbiology of Ghanaian cocoa fermentations analyzed using culturedependent and culture-independent methods. Int. J. Food Microbiol. 114, 168–186. Palles, T., Beresford, T., Condon, S., Cogan, T.M., 1998. Citrate metabolism in Lactobacillus casei and Lactobacillus plantarum. J. Appl. Microbiol. 85, 147–154. Papalexandratou, Z., Falony, G., Romanens, E., Jimenez, J.C., Amores, F., Daniel, H.M., De Vuyst, L., 2011a. Species diversity, community dynamics, and metabolite kinetics of the microbiota associated with traditional Ecuadorian spontaneous cocoa bean fermentations. Appl. Environ. Microbiol. 77, 7698–7714. Papalexandratou, Z., Vrancken, G., De Bruyne, K., Vandamme, P., De Vuyst, L., 2011b. Spontaneous organic cocoa bean box fermentations in Brazil are characterized by a restricted species diversity of lactic acid bacteria and acetic acid bacteria. Food Microbiol. 28, 1326–1338.

87

Payne, M.J., Hurst, W.J., Miller, K.B., Rank, C., Stuart, D.A., 2010. Impact of fermentation, drying, roasting, and dutch processing on epicatechin and catechin content of cacao beans and cocoa ingredients. J. Agric. Food Chem. 58, 10518–10527. Pereira, G.V.D., Miguel, M., Ramos, C.L., Schwan, R.F., 2012. 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. 78, 5395–5405. Pereira, G.V.D., Magalhaes, K.T., de Almeida, E.G., Coelho, I.D., Schwan, R.F., 2013. 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. 161, 121–133. Pettipher, G.L., 1986. An improved method for the extraction and quantitation of anthocyanins in cocoa beans and its use as an index of the degree of fermentation. J. Sci. Food Agric. 37, 289–296. Pipitone, L., 2012. Situation and prospects for cocoa supply & demand. World Cocoa Conference. Intenational Cocoa Organization, Ivory Coast. Quesnel, V.C., 1965. Agents inducing the death of cacao seeds during fermentation. J. Sci. Food Agric. 16, 441–447. Redgwell, R.J., Trovato, V., Curti, D., 2003. Cocoa bean carbohydrates: roasting-induced changes and polymer interactions. Food Chem. 80, 511–516. Reineccius, G., Keely, B.J., Weissberger, W., 1972. Factors affecting the concentration of pyrazines in cocoa beans. J. Agric. Food Chem. 20, 202–206. Rodriguez-Campos, J., Escalona-Buendia, H.B., Orozco-Avila, I., Lugo-Cervantes, E., Jaramillo-Flores, M.E., 2011. Dynamics of volatile and non-volatile compounds in cocoa (Theobroma cacao L.) during fermentation and drying processes using principal components analysis. Food Res. Int. 44, 250–258. Rodriguez-Campos, J., Escalona-Buendia, H.B., Contreras-Ramos, S.M., Orozco-Avila, I., Jaramillo-Flores, E., Lugo-Cervantes, E., 2012. Effect of fermentation time and drying temperature on volatile compounds in cocoa. Food Chem. 132, 277–288. Roelofsen, P.A., 1958. Fermentation, drying and storage of cacao beans. Adv. Food Res. Int. 8, 225–296. Roelofsen, P.A., Giesberger, G., 1947. Onderzoekingen over cacao bereiding. Arch. Koffiecult. Ned.-Indie 16, 119–159. Rohan, T.A., Stewart, T., 1967a. Precursors of chocolate aroma: production of reducing sugars during fermentation of cocoa beans. J. Food Sci. 32, 399–402. Rohan, T.A., Stewart, T., 1967b. Precursors of chocolate aroma — production of free amino acids during fermentation of cocoa beans. J. Food Sci. 32, 395–398. Rohsius, C., Matissek, R., Lieberei, R., 2006. Free amino acid amounts in raw cocoa from different origins. Eur. Food Res. Technol. 222, 432–438. Ruiz, A., Poblet, M., Mas, A., Guillamon, J.M., 2000. Identification of acetic acid bacteria by RFLP of PCR-amplified 16S rDNA and 16S–23S rDNA intergenic spacer. Int. J. Syst. Evol. Microbiol. 50, 1981–1987. Saha, B.C., Racine, F.M., 2011. Biotechnological production of mannitol and its applications. Appl. Microbiol. Biotechnol. 89, 879–891. Saltini, R., Akkerman, R., Frosch, S., 2013. Optimizing chocolate production through traceability: a review of the influence of farming practices on cocoa bean quality. Food Control 29, 167–187. Schwan, R.F., 1998. Cocoa fermentations conducted with a defined microbial cocktail inoculum. Appl. Environ. Microbiol. 64, 1477–1483. Schwan, R.F., Wheals, A.E., 2004. The microbiology of cocoa fermentation and its role in chocolate quality. Crit. Rev. Food Sci. Nutr. 44, 205–221. Schwan, R.F., Rose, A.H., Board, R.G., 1995. Microbial fermentation of cocoa beans, with emphasis on enzymatic degradation of the pulp. J. Appl. Bacteriol. 79, S96–S107. Senanayake, M., Jansz, E.R., Buckle, K.A., 1997. Effect of different mixing intervals on the fermentation of cocoa beans. J. Sci. Food Agric. 74, 42–48. Smetankova, J., Hladikova, Z., Valach, F., Zimanova, M., Kohajdova, Z., Greif, G., Greifova, M., 2012. Influence of aerobic and anaerobic conditions on the growth and metabolism of selected strains of Lactobacillus plantarum. Acta Chim. Slov. 5, 204–210. Soetaert, W., Vanhooren, P.T., Vandamme, E.J., 1999. The production of mannitol by fermentation. In: Bucke, C. (Ed.), Carbohydrate biotechnology protocols, vol. 10. Humana Press Inc., USA, pp. 261–275. Swiegers, J.H., Bartowsky, E.J., Henschke, P.A., Pretorius, I.S., 2005. Yeast and bacterial modulation of wine aroma and flavor. Aust. J. Grape Wine Res. 11, 139–173. Thompson, S.S., Miller, K.B., Lopez, A., Camu, N., 2013. Cocoa and coffee, In: Doyle, M.P., Buchanan, R.L. (Eds.), Food Microbiology: Fundamentals and Frontiers, 4th ed. ASM Press, Washington, DC, pp. 881–889. Voigt, J., Biehl, B., Heinrichs, H., Kamaruddin, S., Marsoner, G.G., Hugi, A., 1994a. In-vitro formation of cocoa specific aroma precursors — aroma related peptides generated from cocoa seed protein by cooperation of an aspartic endoprotease and a carboxypeptidase. Food Chem. 49, 173–180. Voigt, J., Voigt, G., Heinrichs, H., Wrann, D., Biehl, B., 1994b. In-vitro studies on the proteolytic formation of the characteristic aroma precursors of fermented cocoa seeds — the significance of endoprotease specificity. Food Chem. 51, 7–14. Wollgast, J., Anklam, E., 2000. Review on polyphenols in Theobroma cacao: changes in composition during the manufacture of chocolate and methodology for identification and quantification. Food Res. Int. 33, 423–447. Wood, G.A.R., Lass, R.A., 1985. Cocoa, 4th ed. Blackwell Science, London UK. Yu, J., Sun, Z.H., Liu, W.J., Zhang, J.C., Sun, T.S., Bao, Q.H., Zhang, H.P., 2009. Rapid identification of lactic acid bacteria isolated from home-made fermented milk in Tibet. J. Gen. Appl. Microbiol. 55, 181–190. Zak, D.L., Keeney, P.G., 1976. Changes in cocoa proteins during ripening of fruit, fermentation, and further processing of cocoa beans. J. Agric. Food Chem. 24, 483–486. Zalan, Z., Hudacek, J., Stetina, J., Chumchalova, J., Halasz, A., 2010. Production of organic acids by Lactobacillus strains in three different media. Eur. Food Res. Technol. 230, 395–404.

The effect of lactic acid bacteria on cocoa bean fermentation.

Development of a quantitative PCR assay for rapid detection of Lactobacillus plantarum and Lactobacillus fermentum in cocoa bean fermentation.

Fermentation studies of stored cocoa beans.

Estimating cocoa bean parameters by FT-NIRS and chemometrics analysis.

Changes in cocoa proteins during ripening of fruit, fermentation, and further processing of cocoa beans.

Comparative studies on fermentation products of cocoa beans.

Fermentation and aerobic metabolism of cellodextrins by yeasts.

Applying meta-pathway analyses through metagenomics to identify the functional properties of the major bacterial communities of a single spontaneous cocoa bean fermentation process sample.

The bean counters are coming!

Selection of thermotolerant yeasts for simultaneous saccharification and fermentation (SSF) of cellulose to ethanol.

Functional Profile Evaluation of Lactobacillus fermentum TCUESC01: A New Potential Probiotic Strain Isolated during Cocoa Fermentation.

Influence of fermentation and drying materials on the contamination of cocoa beans by ochratoxin A.

Are duplicated genes responsible for anthracnose resistance in common bean?

Behavior of Salmonella during fermentation, drying and storage of cocoa beans.

The key to acetate: metabolic fluxes of acetic acid bacteria under cocoa pulp fermentation-simulating conditions.

Pericytes are Essential for Skeletal Muscle Formation.

NADPH Oxidases Are Essential for Macrophage Differentiation.

Functional analysis of lipid metabolism genes in wine yeasts during alcoholic fermentation at low temperature.

Production of folate in oat bran fermentation by yeasts isolated from barley and diverse foods.

Symbiotic factors in Burkholderia essential for establishing an association with the bean bug, Riptortus pedestris.

Characteristics of Saccharomyces cerevisiae yeasts exhibiting rough colonies and pseudohyphal morphology with respect to alcoholic fermentation.

Cocoa butter-like lipid production ability of non-oleaginous and oleaginous yeasts under nitrogen-limited culture conditions.

Identification of predominant yeasts associated with artisan Mexican cocoa fermentations using culture-dependent and culture-independent approaches.

Headspace flavour compounds produced by yeasts and lactobacilli during fermentation of preferments and bread doughs.