Effect of Microbial Fermentation on Nutritional and Safety Value of Cassava-Based Food

Background: Cassava is a potential energy-rich food crops to make different food products in developing countries but limited by a shortage of protein content and the presence of toxic cyanogenic glycosides. Methods: Cassava-teff our fermented with three pure starter cultures of Saccharomyces cerevisiae, Lactobacillus plantarum and Lactobacillus coryneformis. Two different inoculum levels (0.5 and 1.5 ml) were used. 300 g of cassava-teff our were fermented with each of single starter cultures at 24 and 48 h. Results: The analysis of pH, crude protein and cyanide content indicates fermentation samples with Lactobacillus plantarum and Lactobacillus coryneformis for 48 h with 1.5 ml inoculums resulted in highest pH reduction. Similarly, highest reduction cyanide was recorded with 1.5 ml inoculums of Lactobacillus coryneformis and Lactobacillus plantarum, while the least cyanide reduction was recorded in fermentation samples of Saccharomyces cerevisiae at 24 h of 1 ml of inoculum level, but this value is higher when compared upon boiling (47.77 mg/kg). The highest levels of crude protein were observed fermentation samples with 1.5 ml inoculum of Saccharomyces cerevisiae for 48 h. Regarding palatability, the panelist preferred that fermentation sample with 1.5 ml inoculums of Lactobacillus plantarum and Lactobacillus coryneformis at 48 h having the best taste with the score of 4.90 ± 0.17 and 4.87 ± 0.23 respectively over the control, while sample fermented with Saccharomyces cerevisiae was preferred in terms of avor, texture and overall acceptability with score of 4.87 ± 0.23, 4.73 ± 0.11 and 4.67 ± 0.12 respectively over the control. Conclusions: Therefore, microbial fermentation is a promising candidate for improving nutritional and safety value of cassava- based food and suggested as a choice of the processing method, as this method signicantly reduced cyanide for further analysis. The of to retain for The injera and in an airtight [4,11,17) until the palatability test. The non- fermented cassava mash was used as a control for all fermentation experiments.

The growth of these microorganisms which is more rapid than that of higher plants makes them very attractive as higher protein crop. While only one or two-grain crop is grown per year, yeast may be harvested weekly and bacteria may be harvested daily [5,6]. It has the greatest advantage of cyanide reduction and protein enhancement. Despite its importance as a food and feed in of a Woreda of Walaita Zone, SNNPRS, Ethiopia, not much is known about the role of the fermentative microorganisms in cyanide reduction, protein improvement and palatability in locally produced cassava-based food product. Therefore, this study was tried to investigate the contribution of fermentative microorganisms (lactic acid bacteria and yeast) in the reduction of cyanide, protein enhancement and palatability in cassava-based food.

Methods
Samples were collected from Offa Woreda of the Walaita zone which is found in SNNPRS, Ethiopia.

Experimental Design
The experiment was conducted using three selected pure cultures of cassava fermenting microorganisms i.e Lactobacillus plantarum, Lactobacillus coryneformis, and Saccharomyces cerevisiae and, each at 1 ml and 1.5 ml inoculums level, and two fermentation time (24 and 48 h) and the treatments were factorially arranged in completely randomize design (CRD) with three replications. The non-fermented cassava-teff our was used as a control for all fermentation experiments. The two years edible cassava roots (5kgs) on which investigation carried out was collected from local communities of Busha kebele in Offa woreda and directly brought to the laboratory on the same day using containers with ice to keep them before processing. Teff grain (4 kg) was purchased from Arba Minch Town and ground.
Preparation of cassava our was done according to Girma et al. [13], with some modi cation. Undamaged and uniformly matured raw fresh cassava tubers were taken and washed with tap water to remove adhered soil. The tubers were peeled, cut into pieces and dried to reach moisture content less than 12% using oven glassware cabinet drier (Endecotts LTD, London England at a temperature 65 °C for 24 h. The dried pieces were ground into our by electrical grinder (Moulinex, A2424A, France) and kept in a refrigerator at a temperature of 4°C for the further process.
Teff our was prepared as described by Gebrekidan [11]. Four kilograms (4 kg) of teff grain was manually cleaned and milled. The our was kept in an airtight sealed plastic bucket at room temperature for the duration of the analysis.
The blended mixture was prepared by mixing 50% teff our with 50% cassava our for making injera, protein analysis and cyanide determination.

Microbial Culture Preparation
Lactobacillus plantarum, Lactobacillus coryneformis and Saccharomyces cerevisiae previously isolated from fermented milk were obtained from Ethiopian Biodiversity Institute (EBI). Lactobacillus plantarum and Lactobacillus coryneformis apart from being widely used in the preparation of fermented milk have been reported as the predominant strains among the isolates of traditional sour cassava fermentation [14,15]. Similarly, Saccharomyces cerevisiae is known industrially as important yeast used in the production of a variety of fermented foods. Besides, all the strains have no history of pathogenicity.
Lactobacillus plantarum and Lactobacillus coryneformis reactivated using MRS agar medium and propagated using MRS broth medium [26].
After an incubation period of 18 h at 30 °C, each of cell culture was centrifuged at 3500 rpm for 10 minutes. Then the cell pellets were washed with distilled water and resuspended in distilled water repeatedly until the optical density (OD) of cells reached 0.5 by spectrophotometer at 600 nm. Finally, 1 and 1.5 ml of the diluted sample were used as inocula.
Saccharomyces cerevisiae reactivated using YPD agar medium and propagated in YPD broth medium according to Frier et al. [10]. After the incubation period of 18 h at 25 °C cell culture was centrifuged at 3500 rpm for 10 minutes and suspension was discarded. Then the cell pellets were resuspended in distilled water repeatedly until the optical density (OD) of cells reached 0.5 by spectrophotometer at 600nm. Finally, 1 and 1.5ml of the diluted sample were used as inocula.

Fermentation and Injera Preparation
About 300 g of sieved cassava mixed teff our were weighted into 13 separate 3 L plastic jerry can. Three hundred seventy (370 ml) of sterile distilled water was added into each plastic jerry can and stopped. Then 1 and 1.5 ml inocula of each of Lactobacillus plantarum and Lactobacillus coryneformis were inoculated aseptically to the 2 set of 4 separate plastic jerry can each containing 300 g of sterile cassava mush. The inoculation was accompanied by stirring using a sterile glass rod and allowed to ferment for 24 and 48 h at 30 °C, while 1 and 1.5 ml inocula of Saccharomyces cerevisiae spore suspension was added aseptically to the 1 set of 4 separate plastic jerry can each containing 300 g of sterile cassava mash and allowed to ferment for 24 and 48 h at a temperature of 25 °C. Samples were withdrawn at an interval of 24 and 48 h for the determination of pH [16,28].
After fermentation, about 10% of dough was added into boiling water and waited for 2-5 minutes with continuous string and allowed to cool forming the -Absit‖ and added back to the fermenting dough [2]. After thorough mixing, the batter was fermented at room temperature for 2 h. -Absit ensure injera to have proper texture and consistency by enhancing the dough -rising and gas forming process. After fermentation, a portion of it further dried in a shade for 3 days. The dried product then milled to a powder. Finally, the powder was kept in a refrigerator at 4 °C until used for further analysis. The remaining portion of dough was thinned down to a thick batter and poured in a circular manner onto a hot clay griddle which is lightly oiled, then covered (to retain steam), and baked for approximate 2-3 min. The baked injera was then removed and kept in an airtight container [4,11,17) until the palatability test. The non-fermented cassava mash was used as a control for all fermentation experiments.

Boiling
Boiling of raw fresh cassava roots were done according to the methods described by [1,29]. The raw fresh cassava roots were peeled, and placed into stainless steel pan and boiled for about 45 minutes. The cooked cassava roots were crushed and then shade dried for a week to less than 12% moisture content. The dried pieces were ground into our by electrical grinder (Moulinex, A2424A, France), sieved by 40 mesh size (450 μm) stainless steel sieve (W.S. Tyler Co., Member, Ohio, USA), and kept in a refrigerator at a temperature 4 °C for further analysis.

Determination of Moisture Content
The moisture contents of the cassava root our sample were determined according to the Association of Analytical Chemistry [3], in triplicate.
About 5 kg of chopped fresh cassava tuber was weighed and transferred in to previously dried and weighed glass dishes (porcelain crucibles).
The dishes with cassava samples were placed in a thermostatically controlled oven and heated at 65 °C for 24 h to achieve a constant weight. The dishes were dried again for 30 minutes, cool down and weighed. The drying process was allowed to continue until no more weight loss was recorded between two successive readings. The moisture content then determined by difference and expressed as a percentage [18,19,25].

Determination of PH
The pH of the cassava-based food was determined using a digital pH meter (TH009 (I) A, Arab emirate) in triplicate [6]. The pH was measure before fermentation, after 24 and 48 h by pouring 100 ml of slurry from samples into 250 ml beaker.

Determination of Crude Protein Content
For studying the changes in protein content associated with fermented cassava-teff our with Single starter cultures, inoculums level and time of fermentation was estimated by Lowry's method (1951) with minor modi cation. All the samples mentioned above were weighed 100 mg each and was extracted 50 mM Tris HCl buffer (pH 5.7) containing 5 mM MgCl2, 2mM K2HPO4, 1mM EDTA, 5 mM DTT, 2 mM KH2PO4, 5 mM DTT, 2% PVP, 20% glycerol,10mM NaF, 10mM β-mercaptoethanol and 2mM PMSF. After homogenization, the samples were centrifuged at 4°C centigrade for 20 min at 12,000 rpm. The supernatant was taken and soluble protein content was determined. Bovine serum albumin (BSA Sigma) was used as a standard protein (5-50 μg/ml); absorbance of the samples was recorded at 750 nm.

Preparation of Test Samples for Analysis of Free Cyanide (as HCN Equivalent)
Residual cyanide levels of the ours of the cassava were determined in triplicate using the alkaline picrate method [9], in the modi cation. Five grams (5 g) of each sample was dissolved in 50 ml distilled water and allowed to stay overnight. The sample was ltered and the ltrate was used for cyanide determination. One milliliter (1 ml) of each sample ltrate and standard cyanide solution was measured into seven test tubes respectively and 4 ml of alkaline picrate solution (obtained by dissolving 1 g of picrate and 5 g of Na 2 Co 3 in 200 ml of distilled water) was added to each and the whole set up was incubated in a water bath at a temperature of 50 °C for 5 minutes. After color development (dark red), the absorbance of each content in the test tubes was taken in to a spectrophotometer at 490 nm against a blank containing only 1 ml distilled water Injera produced from the fermented cassava with pure culture and with no culture were subjected to sensory evaluation following the method of Larmond et al. [19]. Panelists were selected from the Arba Minch University postgraduate and undergraduate students. Samples were presented to panelists (4 male and 6 female) in random order during the test day. The sensory attributes were rated on a 5 point hedonic score scale Samples receiving an overall quality score of 3 were considered acceptable.

Statistical Analysis
The analytical determination was made in triplicate. The triplicates per treatment were evaluated for the effect of starter cultures, inoculums level and fermentation time on protein content, cyanide content and consumer acceptance of cassava-based food. The data were analyzed using an Analysis of Variance (ANOVA). Statistical analysis was carried out using SPSS software version 20.0 where possible, mean comparisons were made using the Tukey HSD test at P ≤ 0.05.
Table1. Cyanide and Protein contents in teff our, cassava our and cassava-teff our. There was a signi cance difference (p 0.05) in pH of fermented cassava-teff our due to single starter culture (Table 1,2), size of inoculums (Table 3) and fermentation time (Table 4) except between Lactobacillus plantarum and Lactobacillus coryneformis using the same inoculum level and fermentation time. Cassava-teff our fermented with each of 1.5 ml of Lactobacillus plantarum, Lactobacillus coryneformis and Saccharomyces cerevisiae at 48 h showed pH change from 6.72 ±0.02 to 3.60±0.01, from 6.70± 0.10 to 3.62 ±0.02 and from 6.72 ± 0.02 to 5.12 ±0.10 respectively (Table 3), while using 1 ml inocula of Lactobacillus plantarum, Lactobacillus coryneformis and Saccharomyces cerevisiae pH reduced from 6.66±0.02 to 3.91±0.06, 6.66±0.10 to 3.93±0.05 and from 6.71±0.10 to 5.43±0.11 respectively after 48 h fermentation time. As fermentation time increases, there was also reduction of pH in all starter cultures expect control (non-fermented sample), differing the extent of reduction depending on types of microorganisms, inoculum level and fermentation time. Similarly, Gunawan et al. [14] showed that the optimum pH condition of Lactobacillus plantarum and Saccharomyces cerevisiae was 3.5-4.5 and 3.5-6.0 respectively indicating that cassava fermentation by the action of a single species of micro-organisms can result in a signi cant reduction in pH. The decreases of pH during the fermentation of cassava-teff our results from the production of an organic acid by lactic acid bacteria on the carbohydrate content of cassava root [17,27,28,30].
Table4. pH, protein and cyanide at different times and microbes (inoculum 1.5 ml for all). This shows that further more fermentation of cassava-teff our with Lactobacillus plantarum and Lactobacillus coryneformis for 48 h caused signi cant reduction of cyanide content. The reduced cyanide content of fermented cassava-teff our by Lactobacillus plantarum and Lactobacillus coryneformis were below the safe level recommended by WHO [32]. These ndings are consistent with the results of Kobawila et al. [17], who reported cyanide reduction drastically from 1158 to 339.6 mg/kg after 48 h of fermentation which corresponds to 70.67 % reduction only the difference is initial cyanide content.
Fermentation of cassava our by selected microorganisms result in microbial growth was shown to be essential for the e cient elimination of cyanogens [20,28]. From our investigation, Lactobacillus plantarum and Lactobacillus coryneformis appear to play an important role in cyanide detoxi cation, as already reported by Labri et al. [18] and Tefara et al. [28]. This indicates that it is possible to signi cantly reduce the residual HCN content of cassava through fermentation using Lactobacillus coryneformis and Lactobacillus plantarum. The reduction in cyanide content could be attributed to the ability of the inoculated microorganism (Lactobacillus plantarum and Lactobacillus coryneformis) to produce linamarase which can hydrolyze linamarin and result in degradation of cyanogenic glycosides in to HCN which is subsequently converted in to formamide which is used as both a nitrogen and carbon source [17,23].  [21,22]. The degradation might be due to enzymes linamarase, hydroxynitrile lyase and cyanide hydratase that catalyze the sequential degradation of cyanogenic glycosides into HCN which is subsequently converted into formamide which is used as both a nitrogen and carbon source [23].
Table5. Comparison of the effect of the microbial fermentation using 1.5 ml inoculums at 48 h and boiling on cyanide content of cassava-teff our. % and 51% respectively. This may be in boiling (cooking) enzymatic breakdown of linamarin is small due to heat denaturation of linamarase [7].
Whereas using microbial fermentation was a very e cient process for elimination of cyanide suggests that this method needs to be used for processing of the variety containing a high amount of cyanide [18,24,27].
Crude Protein Single starter cultures, inoculums level and time of fermentation had a signi cant effect (p 0.05) on the crude protein content of fermented cassava-teff our. The protein content of fermented cassava-teff our is signi cantly higher than that of control ( Table 2, 3 and 4). The crude protein content of cassava-teff our fermented with each of 1.5 ml of Lactobacillus plantarum and Lactobacillus coryneformis increased from 4.23± 0.03 % to 6.98 ± 0.01 and 7.02± 0.01 % respectively after 48 h. The extent of increasing protein content depends on types of microorganism, inoculum level and fermentation time which is consistent with the earlier report by Tefera et al. [28] that fermentation of cassavabased food by Lactobacillus Plantarum could have increased protein content up to 4.31%. Okafor et al. [23] also given his observation that cassava mash fermented by Lactobacillus coryneformis increased lysine content 1.2 to 2.45 % after 48 h of fermentation.
The difference is only initial protein content in cassava root that it would appear that the organisms may de nitely play some role in increasing the protein content of fermented cassava teff our because the protein content in the control (non-fermented) cassava-teff our was consistently lower than the cassava-teff our inoculated with Lactobacillus plantarum and Lactobacillus coryneformis. The increase in protein content of fermented cassava-teff our may be because of some microorganisms which degrade cassava pulp by readily could have secreted some extracellular enzymes (proteins) in the cassava pulp [21]. Growth and proliferation of bacteria during fermentation time in the form of single-cell proteins increases that may be possibly accounted for the apparent increase in protein content [5].
The protein content in cassava-teff our fermented with 1.5 ml inoculums of Saccharomyces cerevisiae (13.31±0.02 %) was higher than that of Lactobacillus plantarum and Lactobacillus coryneformis (Table 4). In our observation, further fermentation of cassava-teff our with Saccharomyces cerevisiae for 48 h caused a signi cant (p 0.05) increase in the protein content. The crude protein content of fermented cassava-teff our by Saccharomyces cerevisiae showed in Table 3 and 4 was lower than that reported by Boonnop et al. [5] who demonstrated that fermentation of cassava chips with Saccharomyces cerevisiae could increase crude protein content from 2 % to 32.4 %. The difference could probably attribute to the size of inoculums used and fermentation time. Similarly, the increase protein content also agrees with earlier reports [23] that fermentation of cassava with Saccharomyces cerevisiae would increase protein content, indicating that Saccharomyces cerevisiae had the highest capability to enrich the crude protein content of cassava products.
The increase in protein content in fermented cassava-teff our could be attributed to the ability of Saccharomyces cerevisiae to secret some extracellular enzymes such as amylases, linamarase, and cellulase into cassava mash during their metabolic activities which could lead to yeast growth [5]. This high protein cassava product could very well serve as a protein source in animal diets provided it is economically viable.
Table6. Sensory acceptability scores of injera at different times and microbes (using 1 ml inoculum for all).

Sensory Evaluation of Cassava Based Food (Injera) Flavor
The sensory acceptability showed signi cance (p 0.05) differences between single starter culture and control on the avor of produced injera.
Analysis of variance also had shown signi cance (p 0.05) differences among single starter cultures (Table 6) at different fermentation time (Table 7) and inoculums level (Table 8) except between Lactobacillus plantarum and Lactobacillus coryneformis. The means scores of the avor test of samples are appeared to improve with a longer period of fermentation of cassava-teff our for each starter culture, the highest values being attained around 48 h of fermentation time. The score given to avor acceptability were highest for injera produced from 1.5 ml inoculum of Saccharomyces cerevisiae after 48 h fermentation time, which was 4.87±0.23. Regarding the inoculum level, the highest score observed in injera produced from a 1.5 ml inoculum level for all starter cultures.
The avor acceptability score of the control (none fermented) 2.33 ± 0.29 is lower when compared with injera produced from the fermented cassava-teff our. The microbial activities which increased as fermentation continued might have accounted for the perceived differences in the avor of the product fermented for different lengths of fermentation time. The avor of food depends on the balance of volatile compounds those produced during fermentation. A vast number of volatile compounds may be synthesized and modulated by Saccharomyces cerevisiae during fermentation. In line with this nding, Tefera et al. [28] reported that Saccharomyces cerevisiae was able to produce compounds such as organic acids, alcohols, aldehydes, and carbonyls which have imparted appealing avor to the fermenting cassava.
A previous study by Hasan et al. [15] on fermented rice also showed an increase in volatile compounds due to fermentation by yeasts and lactic acid bacteria. Similar to the current study it has been reported that Saccharomyces cerevisiae contributes to avor development while fermenting rice for injera production.

Color and Taste
Regarding Color of cassava-based food (injera) the highest score was 4.83 ± 0.28 and 4.67 ± 0.29 by 1.5 and a 1ml inoculum of Saccharomyces cerevisiae respectively at 48 h fermentation time, while the least score 4.17±0.29 was for control. However, Analysis of variance showed that using a single starter culture ( Table 6), time of fermentation ( Table 7) and addition of inoculum level (Table 8)  The sensory score of taste appears greater than three in all cassava-based food (injera) produced by fermentation except for control (nonfermented). Analysis of variance showed that using single starter culture, addition of inoculums level and fermentation time had shown signi cant (p 0.05) difference regarding taste attributes of cassava-based food, but no signi cant difference (p 0.05) was detected between Lactobacillus plantarum and Lactobacillus coryneformis on taste attributes of cassava-based food ( Table 6,7 and 8). This indicates that all organisms seem to be responsible for the taste of the cassava-based food as suggested by Odibo and Umeh [22] and con rmed by the present study. However, panelists rated the sample fermented with 1.5 ml inoculums of Lactobacillus plantarum and Lactobacillus coryneformis at 48 h as having the best taste with the score of 4.90± 0.17 (98%) and 4.87 ± 0.23 (97%) respectively, but the sample fermented with 1 ml inoculums of Saccharomyces cerevisiae at 24 h fermentation time as having lower taste with the score of 3.13±0.23 (62.6%).
As the fermentation time increased, the scores for taste also increased in each starter culture. This result agrees with earlier reports by Tefera et al. [28] that fermentation of cassava-based food (chike) by Lactobacillus mesenteroides result in 72% score of taste, only a difference is inoculum level used. This might be possibly attributed to the fact that Lactobacillus plantarum and Lactobacillus coryneformis converts the sugars in fermenting substrate (primarily glucose and fructose) to lactic acid, acetic acid, ethanol, CO2 and other avor compounds [24,25,31].

Texture
As shown in Table 6, 7 and 8, there was a signi cant difference (p 0.05) in the texture acceptability scores between control and starter cultures as well as among starter cultures, inoculum levels and fermentation time of cassava-based food. However, the least scores 3.07±0.4 and 3.07±0.06 were for the cassava-based food fermented with 1 ml inoculums level of Lactobacillus plantarum and Lactobacillus coryneformis respectively at 24 h. The highest mean score 4.73±0.11 was observed for the cassava-based food fermented with 1.5 ml Saccharomyces cerevisiae at 48 h fermentation time followed by 1.5 ml Lactobacillus plantarum (4.17±0.47) and Lactobacillus coryneformis (4.17±0.29). This showed that the starter culture, inoculums level and fermentation time in uences the quality of dough thereby that of the texture of the injera. This indicates that Saccharomyces cerevisiae possesses cellular activities, where also found to be contributed to the modi cation of cassava texture during fermentation [21,29].
Table8. Sensory acceptability scores of injera using different inoculum size and microbes (time 48 h for all).