Controlled fermentation of the zoom-koom dough using two isolates of lactic acid bacteria (LAB 1 and LAB 5) as starter cultures: Effect on hygienic, rheological, nutritional and sensorial characteristics of the final product

Zoom-koom is a popular non-alcoholic beverage in Burkina Faso, which is based on cereals and mainly produced by women with important socio-economic implications. This study aimed to evaluate the effect of controlled fermentation using two selected isolates of lactic acid bacteria (LAB) as starter cultures, on the rheological and hygienic quality of zoom-koom. The starter cultures were used singly in monoculture and both in mixed culture. Microorganisms dynamic during the controlled fermentation were followed and enumerated using pour plate methods. The titratable acidity, pH, viscosity, water, ash, crude protein (N×6.25), crude fat and total carbohydrates contents were determined on the final zoom-koom by using standards methods. Sensory analyses of zoom-koom samples were performed by a panel of 30 tasters. The enterobacteria counts of all the controlled fermented zoom-koom samples using starters cultures decreased totally and significantly (p˂0.05) from 6.4 (LAB 1), 5.5 log CFU/g (LAB 5) and 3.8 (LAB 1 and 5 in mixed culture) to ˂ 1 log CFU/g after 24 h of fermentation. However, those of natural fermentation without inoculum decreased significantly (p˂0.05) but not totally (1.4 log CFU/g after 24 h of fermentation). The z oom-koom from LAB 5 presented the best production of exopolysaccharides and was more viscous and homogenous than the others. All the zoom-koom samples presented a low fat (4.74, 5.21, 5.36 and 5.55%/DM) and ash (0.32, 0.53, 0.49 and 0.69%/DM) contents with a high total carbohydrate (74.18, 76.24, 68.86 and 76.09%/DM) and protein (20.75, 18.02, 25.30 and 17.66%/DM) contents. The most appreciated zoom-koom by the tasters was the controlled fermented zoom-koom from mixed culture (LAB 1 and 5).


INTRODUCTION
For Africans, the importance of traditional food fermentation lies in providing improved flavors to existing staples (for example cereals and root crops), and as a cheap way for food preservation and enhancement of the nutritional quality and digestibility of the raw products (Olasupo et al., 2010). Frequently, fermented foods are considered to have health benefits, and in many regions, they are believed to aid in the control of some diseases, in particular intestinal disorders (Mathara et al., 2004). Traditional fermented foods still play a major role in the diet of numerous societies worldwide. The African dietary ethos includes both fermented and unfermented cereals and cassava products, wild legume seeds, but also meat, milk products and alcoholic beverages (Tamang and Samuel, 2010). Zoom-koom is one of common streetvended beverage and it is produced by crafts women. It is sold in all parts of Burkina Faso, mainly in cities such as Ouagadougou, Bobo-Dioulasso and Koudougou (Icard-Vernière et al., 2010). The grains of millet or sorghum are soaked overnight and then washed and mixed with spice (ginger and mint). The blend is ground into a dough, diluted with water, and then filtered using a clean muslin cloth to obtain zoom-koom, in which sugar and tamarind juice are added to give a sweet and sour taste. The production of zoom-koom is usually done in unhygienic environmental conditions (Besadjo-Tchamba et al., 2014;Soma, 2014;Tapsoba et al., 2017a).
Recently, study on the traditional process of zoomkoom, showed the positive impact of the fermentation on the hygienic quality of this drink (Tapsoba et al., 2017a). Some isolates of lactic acid bacteria (LAB) involved in the zoom-koom production process identified as Weissella cibaria/confusa had shown their ability to produce exopolysaccharides and antimicrobial compounds (Tapsoba et al., 2017b). These technological properties are very important for the improvement of the safety and the texture of the zoom-koom in controlled fermentation. For example, the use of exopolysaccharides (EPS)producing LAB strains as ferment during the production of fermented milks improved the texture and decreased the syneresis (Zannini et al., 2016). The success of EPS application in the food industry is generally dictated by its ability to bind water, interact with proteins, and increase the viscosity of the milk serum phase. EPS may act as texturisers and stabilisers, and consequently, avoid the use of food additives (Duboc and Mollet, 2001;Zannini et al., 2016). The availability of LAB starter cultures to produce exopolysaccharides in situ during fermentation could be a suitable alternative for products whose polysaccharides addition requires the specification as food additives, which is a condition not much appreciated by consumer. Zoom-koom is a suspension of millet fermented dough in water, which settles quickly. The use of EPS-producing LAB isolates, for controlled fermentation could improve the physical stability of this beverage. Moreover, LAB are generally recognized as safe (GRAS) due to their long history of safe use in food  production, and many of them have the qualified presumption of safety (QPS) status (Lahtinen et al., 2011;Caggianiello et al., 2016). Controlled fermentation using starter cultures allowed improvement of the hygienic and nutritional quality of traditional fermented products (Egounlety et al., 2007;Sawadogo-Lingani et al., 2008;Yao et al., 2009;Soma, 2014). This study aimed to use two isolates of LAB producing EPSs and antimicrobial compounds (LAB 1 and 5) as starter cultures, to improve the rheological, nutritional, sensory and hygienic quality of zoom-koom.

Origin of starters' cultures
The LAB isolates (LAB 1 and 5) used as starters cultures (pure cultures) were obtained from traditional fermentation process of zoom-koom (Tapsoba et al., 2017a). These isolates were previously characterized and identified as W. confusa/cibaria by using 16S rRNA gene sequencing and were able to produce EPSs and antimicrobial compounds against Escherichia coli; Pseudomonas aeruginosa and Salmonella thyphimerium (Tapsoba et al., 2017b).

Preparation of LAB inoculums
The two selected LAB isolates (previously stored in MRS-broth + glycerol at -20°C) were subcultured onto mMRS agar and incubated for 48 h at 37°C. The isolated colonies were then subcultured in 10 mL of MRS-broth and incubated for 24 h at 37°C. 0.1 mL of culture broth of each tube initially prepared was subcultured in MRS-broth (10 mL) and then incubated for 16 to 18 h at 37°C. For each isolate, the culture broth obtained after 16-18 h of incubation was distributed in sterile cryotubes (1 mL/tube) then centrifuged at 5000 g for 10 min. The supernatant of each tube was removed and the pellet (cells) of the tube was retained. To this pellet was added 1 mL of sterile diluent [0.1% (w/v) peptone (Difco), 0.85% (w/v) NaCl (Sigma), pH 7.2 ± 0.2] after vortexing, a further centrifugation was carried out at 5000 g for 10 min. The supernatant was again removed and the pellet was kept. One millimeter (1 mL) of sterile diluent was added to the pellet and, after stirring, the suspension of cells which constitutes the inoculum was stored in the refrigerator at 4°C. The concentration of viable cells of the inoculum was determined by enumeration on mMRS agar. The inoculum was used at a rate of 1% (v/v) (Sawadogo-Lingani et al., 2008;Soma, 2014) in the millet dough for controlled fermentation.

Controlled fermentation using the isolates
Controlled fermentation in monoculture was carried out at 30°C in an incubator (Binder 78532 Tuttlingen, GERMANY) using separately LAB 1 and 5. For each isolate, 20 mL of inoculums were prepared to inoculate 2 L of millet dough made with millet. For controlled fermentation in mixed culture (with both LAB 1 and 5), 20 mL of mixed inoculum (10 mL of LAB 1 inoculum + 10 mL of LAB 5 inoculum) was used to inoculate 2 L of millet dough. The controlled fermentation of the millet dough with the isolates were followed by   (Tapsoba et al., 2017a). sampling at intervals of: 0, 4, 6, 8, 10 and 24 h for laboratory analyses. For each sample, pH, titratable acidity, mesophillic microorganisms, lactic acid bacteria, enterobacteria, yeasts and molds were measured or counted. A natural fermentation of the millet dough without inoculum was carried out simultaneously to serve as a control at each fermentation trials. For each isolate, the trial fermentation was done in duplicate. The flow diagram ( Figure  1) of zoom-koom previously described (Tapsoba et al., 2017a) was adapted for the production with controlled fermentation.

Millet whole grains
Washing (

Unfermented zoom-koom
Fermented zoom-koom Spain), and incubated at 37°C for 24 h according to ISO 7402 (1993). The results were given as CFU/g or mL of sample. The trial were done in duplicate.

Physico-chemical and nutritional analyses
The pH of the samples was measured with an electronic pH meter (Model HI 8520; Hanna Instrument, Singapore). For solid samples, 10 g of product were mixed with 20 mL of distilled water prior to pH measurement. For liquid samples, the pH was measured directly (Sawadogo-Lingani et al., 2007). For titratable acidity determination, 5 g or 5 mL of sample suspended in 30 mL of ethanol (90°) was mixed 1 h, using an automatic agitator, and centrifuged for 5 min at 3500 g. From the supernatant, 20 ml was transferred to a 50 ml measuring flask and was titrated with NaOH 0.1 N using 1% phenolphthalein as indicator (Soma, 2014). The titratable acidity (as g lactic acid per 100 ml or g of sample) was calculated according to Amoa-Awua et al. (1996). Water content was determined by oven drying the sample at 105 ± 2°C for 12 h (NF V03-707, July 2000); ash content was determined by incineration at 650°C overnight according to the French standard V03-760 (1981); crude protein content (N×6.25) was determined by the Kjeldahl method after acid digestion (NF V03 50, 1970); crude fat content was determined by soxhlet extraction using n-hexane (ISO 659, 1998). Total carbohydrates content were determined by spectrophotometric method at 510 nm using orcinol as reagent (Montreuil and Spik, 1963). The values were expressed in g/100 g of dry matter. The trial were done in duplicate

Determination of viscosity
The viscosity measurement of the zoom-koom samples resulting from the controlled fermentations by the LAB 1 and LAB 5 isolates, was carried out by using a viscometer (CSC scientific 1-800-458-2558). This measure consisted sinking 10 mL of the zoom-koom samples on a viscometer and measuring the flow rate. The result was expressed in cm/s. The types of zoom-koom were left for settling to observe their homogeneity at different times (25 min and 24 h).

Sensory analysis of zoom-koom samples
The sensory analysis consisted of evaluating the sensory profile of zoom-koom samples : A test of differentiation of the controlled fermented zoom-koom samples compared to the unfermented zoom-koom; used as control sample; a test of the classification of the zoom-koom samples according to the tasters were also performed. Thirty (30) members tasting panel were composed of men and women aged between 15 and more, who had already consumed the zoom-koom. The sensory profile was related to the color (nice, acceptable and mediocre), mouthfeel (very pleasant, pleasant and unpleasant), sweetened taste (very sweet, sweet and little sweet), aroma (very good, good and fair) and acidity (very acidic, acidic and fair acidic).

Statistical analysis
All the data (except sensorial analyses data) were subjected to Analysis of Variance (ANOVA) with the statistical software XLSTAT-Pro 7.5.2 and the means were compared using the test of Student Newman-keuls to the probability level p˂0.05. The curves were obtained using Microsoft Excel 2013. The data of sensorial analyses were performed using the Chi 2 test with the statistical software SPSS.

Microbial growth during fermentation
The inoculum counts were 10 6 CFU/mL. All the control samples showed the same trend with their corresponding controlled fermentation trials. In this study, one control was presented to illustrate the other controls.
From the results, it is shown that during all the fermentation trials, the enterobacteria counts decreased significantly (P˂0.05) after 24 h of incubation (Figures 2, 3, 4 and 5). Thus, from the fermentation using LAB 1 and 5 isolates (singly) as starters cultures in monoculture, the enterobacteria counts decreased from 6.4 (0 h) to ˂ 1 log CFU/g (24 h) (LAB 1 in monoculture) and from 5.5 (0 h) to ˂ 1 log CFU/g (24 h) (LAB 5 in monoculture) as shown in Figures 2 and 3. From the fermentation using both isolates LAB 1 and 5 in mixed culture, the enterobacteria counts decreased from 3.8 (0 h) to ˂ 1 log CFU/g (24 h) as shown in Figure 4. The natural fermentation of millet dough (control) also showed a significant decrease (P˂0.05) in enterobacteria counts ( Figure 4). However, the enterobacteria counts at 24 h of fermentation were not ˂ 1 log CFU/g. These counts were 1.3 log CFU/g for the natural fermented millet dough samples at 24 h of fermentation ( Figure 4). All the final products (zoomkoom) did not contain enterobacteria except the natural fermented zoom-koom samples ( Table 1). The yeasts, LAB and mesophillic microorganisms counts increased significantly (P˂0.05) after 24 h of fermentation ( Figures  2, 3, 4 and 5). Thus, the yeasts counts increased from 4.2 (0 h) to 7.1 log CFU/g (LAB 1 in monoculture) as shown in Figure 2, from 5.3 (0h) to 7.2 log CFU/g (LAB 5 in monoculture) as shown in Figure 3 and from 5.0 (0 h) to 6.8 log CFU/g (both LAB 1 and 5 in mixed culture) as shown in Figure 4. The natural fermentation showed the same trend. The LAB counts increased from 8.3 (0 h) to 8.7 log CFU/g (LAB 1 in monoculture), from 7.9 (0 h) to 8.6 log CFU/g (LAB 5 in monocuture) and from 6.7 to 8.7 log CFU/g (both in mixed culture). The natural fermentation showed the same trend. No moulds were observed after 24 h of fermentation for all the fermentation trials. For all the fermentation trials (natural and controlled fermentation) the LAB, mesophillic microorganisms and yeasts counts decreased a little in the final product (zoom-koom) after diluting and filtering of the dough (at 24 h of fermentation) as shown in Table  1.
From the means comparison of all the controlled fermentation trials, the LAB, mesophillic microorganisms and yeasts counts at 24 h of fermentation were significantly different (P˂0.05) from those of 0 h ( Figures  2, 3 and 4). The enterobacteria, mesophillic microorganisms, LAB and yeasts counts of natural fermentation at 24 h of fermentation of the dough, were significantly different (P˂0.05) from those of 0 h as shown in Figure 5.

Physicochemical parameters during fermentation pH
The pH of controlled fermented samples evolved similarly during all the trials fermentation processes. The pH values decreased significantly after 4 h of fermentation (p˂0.001). The pH obtained with the fermentation in mixed culture (LAB 1 and LAB 5) showed the lowest decrease after 4 h of fermentation (from 6.2 to 5.4). The pH decreased slowly (from 6 to 10 h) before stabilizing at pH 4.0 (10 to 24 h) for monoculture fermentation (Figure 6). It should also be noted that the pH measured during the natural fermentation of the millet dough without inoculum (control) showed a similar evolutionary trend as that performed with the LAB 1 and LAB 5 isolates ( Figure 6).

Titratable acidity
The titratable acidity of all the samples showed the same evolutionary trend during the trials fermentations. The results show that titratable acidity evolved significantly from 0 to 24 h for all trials fermentations. The highest acidity value was recorded with the monoculture fermentation using the LAB 1 isolate at 24 h (1.24 g of lactic acid/100 g). After dilution and filtration of the 24 h fermented dough, the titratable acidity values of all the fermentations decreased significantly (p˂0.001) as shown in Figure 10. It should also be noted that the titratable acidity measured during natural fermentation of the millet dough without inoculum (control) showed a similar evolutionary trend as that performed with the LAB 1 and LAB 5 isolates (Figure 7).

Viscosity
From the results of viscosity, the flow tests showed that the zoom-koom fermented by the isolate LAB 5 was the most viscous and homogeneous with a flow of 0.22 cm/s as compared to the natural fermented zoom-koom (control for LAB 5) without inoculum (0.14 cm/s). The flow of the other types of zoom-koom fermented in monoculture with the isolate LAB 1 and in mixed culture with the isolates both LAB 1 and 5 were different from that of the control and less homogeneous than the zoomkoom with the isolate LAB 5. The unfermented zoomkoom was the least viscous and decanted faster than other fermented types. After 25 min and 24 h of settling, the control (natural fermented) zoom-koom settled faster than the controlled fermented zoom-koom using LAB 5.   The last one was more viscous and cloudy.

Nutritional characteristics of fermented zoom-koom samples
The fermented zoom-koom sample from monoculture with the isolate LAB 5 contained less water and more dry matter content than the others, but not significantly different on statistical plan (p˂0.05) ( Table 2). This sample contained more fat, total carbohydrates and ash than the zoom-koom sample with isolate LAB 1. However, the zoom-koom sample with isolate LAB 1 contained more proteins as compared to the zoom-koom with isolate LAB 5. Both samples contained more sugars than the mixed culture fermentation sample. The control sample (natural fermentation without inoculum) contained more fat and ash than the others. The highest ash contents were obtained with the zoom-koom LAB 5 samples and the natural fermented zoom-koom sample  without inoculum (control). Natural fermented zoom-koom (without inoculum) showed the best fat levels and the lowest value of proteins. All the samples showed a low level of fat. The highest protein content was obtained with the zoom-koom sample from fermentation in mixedculture (Table 2), probably due to the high contribution of isolate LAB 1. No significant difference (p˂0.05) was observed for water content. The protein content of the zoom-koom sample from the fermentation in mixedculture was significantly different from that of the other samples (p˂0.05). The ash contents of the different samples were significantly different from each other (p˂0.05). The total carbohydrates in the zoom-koom from mixed culture fermentation were significantly different from the others (p˂0.05).

Sensorial characteristics of fermented zoom-koom samples
From the sensory analysis results, it appeared that 70% of the tasters found that the zoom-koom resulting from the monoculture fermentation with the isolate LAB 1 and the zoom-koom resulting from the fermentation in mixed culture with both isolates LAB 1 and LAB 5 had a nice color. However, 50 and 13.3% of the tasters found that the zoom-koom from the monoculture fermentation with the isolate LAB 5 and the control zoom-koom showed a nice color (Figure 8). The zoom-koom with the isolate LAB 1 and the zoom-koom from the fermentation in mixed-culture showed a better aroma (46.7 and 46.7% of the tasters, respectively) than the zoom-koom with the isolate LAB 5 and the control zoom-koom (40 and 43.3%, respectively) according to the tasters (Figure 8). The mouth feel after tasting the zoom-koom in mixed culture (both LAB 1 and Lab 5) and the control zoom-koom appeared pleasant (63.3 and 63.3%, respectively). Approximately 60% of the tasters appreciated pleasant mouth feel after tasting the zoom-koom resulting from monoculture fermentation with isolate LAB 1 on one hand and isolate LAB 5 on the other hand. The tasters (73.3%) also found that the zoom-koom with isolate LAB 1 and the control zoom-koom were sweet. However, 60% of the tasters found that the zoom-koom resulting from the monoculture fermentation with the isolate LAB 1 was sweet, while 70% of the tasters found that the zoomkoom in mixed culture (both LAB 1 and LAB 5) was sweet ( Figure 8). The control zoom-koom for sensory analysis is an unfermented zoom-koom.
As for the acidity, 70, 66.7 and 60% of the tasters found that zoom-koom with LAB 1, LAB 5 isolate in monoculture and mixed culture, respectively had normal acidity against 13,3% of tasters who thought that the control zoom-koom was fairly acid (Figure 9). A differentiation test based on the homogeneity and texture of the different types of zoom-koom was performed as compared to the control zoom-koom. From this, it appeared that all the fermented zoom-koom were different from the control zoom-koom according to the tasters. Also, 83.34% of the tasters found that the control zoomkoom sample had more liquid (less viscous) than the controlled fermented zoom-koom samples. Among the three controlled fermented zoom-koom samples, the zoom-koom obtained with isolate LAB 5 was the most viscous and cloudy according to the tasters. All tasters did not notice any pungent taste of ginger in all zoomkoom samples after tasting. The overall ranking of tasters were the zoom-koom resulting from mixed-culture fermentation at the first place (33.3%), followed by zoom-koom with isolate LAB 1 in second place (36.7%), zoom-koom with isolate LAB 5 was in third place (36.7%) and the zoom-koom control was fourth (63.3%) according to the proportion of tasters for each rank (Figure 10).

DISCUSSION
The enterobacteria, mesophillic microorganisms and LAB counts increased after 4 h of fermentation, while the yeast counts remained almost unchanged. This increase could be due to the fact that at the beginning of the fermentation, the medium was rich in nutrients with a favorable temperature which allowed the growth of microorganisms. Indeed, water activity and the presence of nutrients could promote the activation of spores, the growth of bacteria, yeasts and molds (Tawaba et al., 2013). LAB are generally described as mesophilic microorganisms with an optimal growth temperature of 30°C (van de Guchte et al., 2002). Gymnase (2011) also indicated that cereals contain prebiotics which stimulate the growth of bacteria like enterobacteria and LAB for the present study and zoom-koom is a cereal based beverage. The decrease in enterobacteria counts during the fermentation (8, 10 and 24 h) is probably due to the growth of LAB which are well known to produce antimicrobial substances such as organic acids (lactic, acetic, formic and caproic phenolic), carbon dioxide, hydrogen peroxide, ethanol and bacteriocins during fermentation (Messens and De Vuyst, 2002).
The pH of the fermented dough remained stable at pH 4.0 from 10 to 24 h and this induced an effective action of the acidity on enterobacteria. The results corroborate those of Soma (2014) who observed a decrease in enterobacteria counts in the fresh zoom-koom after 24 h of fermentation, using a strain of Lactobacillus fermentum as starter. This result also confirmed previous study of Tapsoba et al. (2017a)   enterobacteria and yeasts counts was found after 10 h of natural fermentation of the zoom-koom based on millet and red sorghum dough. In addition, the LAB isolates used as starter in this study have been selected on the bases of their antibacterial and antifungal activities (Tapsoba et al., 2017b). The presence of enterobacteria in the control dough at 24 h and their absence in the dough with the inocula, means that the selected isolates have maintained and expressed their antibacterial properties. In fact, these selected isolates were able to produce bacteriocins like compounds or similar metabolites according to the antimicrobial activities performed (Tapsoba et al., 2017b). It was also observed in previous study that the bacteriocin produced by W. confusa had a broad spectrum of antimicrobial activity inhibiting both Grampositive and negative bacteria (Hweh and Koshy, 2015). The results also highlighted an increase of yeasts population after 24 h of fermentation, while those of LAB decreased; this could be due to the fact that LAB by their Zoom-koom in mixed culture (LAB 1 and LAB 5) Control zoom-koom