IV. Plant derivatives

12 Cassava Processing in Africa

Olusola B. Oyewole

Cassava is an important food crop in the tropics and many countries in Africa. The crop contributes significantly to the diets of over 800 million people, with per capita consumption averaging 102 kilograms per year. In some areas of Africa it constitutes over 50 percent of the daily diets of the people.

Traditionally, cassava is processed before consumption. Processing is necessary for several reasons. First, it serves as a means of removing or reducing the potentially toxic cyanogenic glucosides present in fresh cassava. Second, it serves as a means of preservation. Third, processing yields products that have different characteristics, which creates variety in cassava diets.

The objective of this paper is to detail the strategy and program being followed in our laboratory to utilize the knowledge of biotechnology to improve the processing of cassava in Africa.

TRADITIONAL PROCESSING

Processing of cassava for food involves combinations of fermentation, drying, and cooking. Fermentation is an important method common in most processings. While there are many fermentation techniques for cassava, they can be broadly categorized into solid-state fermentation and submerged fermentation. Solid-state fermentation, typified by gari production, uses grated or sliced cassava pieces that are allowed to ferment while exposed to the natural atmosphere or pressed in a bag. Submerged fermentation involves the soaking of whole peeled, cut and peeled, or unpeeled cassava roots in water for various periods, as typified by the production of fufu and lafun in Nigeria. Traditionally, cassava is fermented for 4 to 6 days in order to effect sufficient detoxification of the roots.

(The support of the International Foundation for Science, Stockholm, Sweden, is gratefully acknowledged. )

Some processors, out of economic pressure, ferment cassava for less than 2 days. Some cases of food poisioning have been attributed to this practice. Application of biotechnology to traditional cassava processing has prospects for producing safe and well-detoxified products.

RESEARCH APPROACH

Our approach on cassava processing research is divided into three areas:

1. Investigating the science of the traditional submerged fermentation of cassava to fufu and lafun production;

2. Optimization of the processing through process controls; and

3. Improvement of traditional processing through application or biotechnological techniques.

The microorganisms involved in the submerged fermentation process have been isolated and found to include Bacillus subtilis, Klebsiella sp., Candida tropicalis, C. krusei, and a wide spectrum of lactic acid bacteria, major among which are Lactobacillus plantarum and Leuconostoc mesenteroides. A microbial succession trend was found with the starch degrading Bacillus subtilis, giving way to the lactic acid bacteria and yeasts that dominate the latter part of the fermentation. The pH and titratable acidity of the fermenting cassava roots increased from 6.3+0.2, 0.08+0.03 percent to 4.0+0.3, 0.36+0.05 percent, respectively, after the 96-hour fermentation period. Organic acids produced during fermentation include lactic, acetic, propanoic, and butanoic acids, among others, and these are believed to contribute to the characteristic flavor of fermented cassava products. Fermentation causes release of some bound minerals, including calcium and magnesium. The most important contribution of fermentation is the release from the plant tissues of the enzyme linamarase, which is involved in the breakdown of the linamarin and lotaustralin (cyanogenic glucosides) of cassava, which releases hydrogen cyanide and thus detoxifies the product.

PROCESS OPTIMIZATION

Processing conditions for optimizing the fermentation process have been investigated in our laboratory. We found that the temperature range of 30° to 35°C, with a soaking period of 48 to 60 hours, is best for submerged processing. The size to which the roots are cut, peeling or nonpeeling of roots before processing, changing or nonchanging of fermentation water at intervals during processing, and the age of roots all affect the characteristics of the final product. In addition, the protein contents of products can be improved by cofermentation with legumes such as soybeans and cowpea.

BIOTECHNOLOGICAL INVESTIGATIONS

The overall goal of our biotechnological investigations is to develop an appropriate starter culture for cassava processing that will effectively produce linamarase enzymes for detoxifying cassava, break down starch to the simple sugars needed for acid production, improve the protein content of the products, reduce processing time, and yield products with stable desired qualities. The following summarizes our current findings:

· The microorganisms involved in fermentation have been characterized.

· Characterized isolates were used as single and multiple starter cultures for cassava fermentation. This has made it possible to understand the roles of each of the microorganisms implicated in the natural fermentation process. Bacillus subtilis and Klebsiella spp. contribute significantly to the rotting of cassava roots. In addition, B. subtilis produces amylase enzymes that are necessary for the breakdown of starch to sugars, which are needed for the growth of other fermenting microorganisms, including the tactics. Yeasts play a major role in odor development and, where high yeast biomass is encouraged, protein-enriched products are not. Lactic acid bacteria convert cassava sugars to lactic and other acids that contribute to the flavor in addition to having preservative effects.

· Appropriate starters have been developed that can produce amylase and linamarase enzymes necessary for starch breakdown and cyanogenic glucoside hydrolysis; two major biochemical processes needed in cassava processing. For this, the lactic acid bacteria were investigated since they were the predominant microbial group present at the beginning of fermentation and which persist and survive the acidic conditions that prevail in cassava fermentation. To date, we have found strains of Lactobacillus plantarum that are capable of producing amylase and linamarin. The linamarase produced has been purified, and it exhibits optimal activity at pH 5 to 8 and temperature of 30° to 40°C. Prospects for cassava processing using a selected single culture with properties for starch hydrolysis, cyanide detoxification, and acid production have thus evolved.

· To initiate genetic manipulation of cassava lactic acid bacteria, the plasmid profiles of the lactobacilli isolated from cassava were studied. The presence of plasmids among cassava lactobacilli has been confirmed. Further research is needed to investigate the correlation between possession of plasmids and linamarase production in order to establish prospects for genetic manipulation.

FUTURE RESEARCH

Beyond the selection of appropriate starter cultures for cassava fermentation, it will be necessary to improve the starter culture. Genetic manipulation of the starter culture offers the best hope for improved cassava processing, with higher economic returns and improved stable qualities.

Cassava processing could also be enhanced by using biotechnological principles to modify structural and processing characteristics of cassava cultivars to meet specific product requirements.

The linamarase elaborated by cassava plant tissues and fermenting microorganisms has been found to be unstable under high acidic conditions characteristic of the latter part of natural fermentation. Techniques for increasing the stability of linamarase enzyme to acidic conditions could be investigated.

The usefulness of cassava fermenting microorganisms could be further investigated for the production of other economically viable products such as acidulants and antimicrobial agents.

A biotechnological approach could be investigated for the treatment of odorous fermented cassava water and cassava root peels.

13 Improving the Nutritional Quality of Ogi and Gari

T. G. Sokari

Ogi is a blancmange-like product processed by fermenting the slurry from wet-milled maize (or sorghum or millet). Used as both a weaning food for infants and as a breakfast food by adults, ogi is one of the most important food items in Nigeria. Yet it is nutritionally inferior to maize, which is deficient in certain essential amino acids, because of the maize-milling process that is an integral part of ogi production.

Cassava, another very important food crop, has the problem of possible nutritional complications because it contains the cyanogenic glucosides linamarin and lotanstralin. Although the cyanogens in cassava are hydrolyzed to hydrogen cyanide during processing by the endogenous enzyme linamarase (1,2), not all processes are equally effective. It has even been suggested that traditional processing techniques are unlikely to remove all the cyanide from cassava (3,4).

In view of this, studies were undertaken to increase the protein content of ogi relatively inexpensively and to develop a technique for processing cassava into gari that would eliminate cyanogens from the product or reduce them to innocuous levels.

PROCESSING OF OGI

The traditional technique for processing maize into ogi is summarized in Figure 1. Also shown in Figure I is an alternative to this procedure; a 20-minute boiling step is substituted for the normal 24- to 28-hour steeping of maize prior to wet milling (5). Cowpea can be combined with maize to increase the protein content of ogi.

Substituting 20 minutes of boiling for the traditional 24 to 28 hours of steeping prior to wet milling maize reduced processing time from 72-76 hours to about 24 hours.

FIGURE 1 Processing of maize into ogi including cowpea protein enrichment; dotted lines show a bypass by which processing time can be reduced.

There was, however, no significant difference ( p> 0.05) in the aroma, color, taste, and overall acceptability between the products obtained by the short-time processing and traditional processing (5). The same was also true for unenriched and protein-enriched ogi except for color (6).

CYANIDE REDUCTION DURING CASSAVA PROCESSING

Two foods processed from cassava (gari and ijapu) were studied. Adding water to grated cassava at the 75 percent (v/w) level and heating at 50°C for 6 hours resulted in linamarin reduction of >99 percent (Figure 2). The pH of the mash fell from 6.4 to 6.3 during the period (7). After dewatering, the mash was adjusted to a pH below 4 by equilibrating with a 3-day fermented cassava liquor (40 percent, v/w) at 50°C for 12 to 18 hours. The equilibrated mash was then dewatered and toasted (Figure 3). A panel of tasters who were familiar with gari but otherwise untrained could not differentiate between the product and traditionally processed gari. Both sets of products were equally acceptable to the panel.

Figure 2 Effect of added water at 0% - i.e., control o-o, 25% D-D, 50% v-v, 75% - levels on linamarin hydrolysis in grated cassava at (a) 30°C, (b) 40°C, (c) 50°C.

FIGURE FIGURE 3 Processing of ijapu and gari from cassava.

The modified procedure for processing cassava into gari reduced the processing time from >96 hours to about 24 hours. Cyanogens were not detectable in the product by the method of Cooke (8,9).

The study of the traditional production of ijapu (Figure 3) was intended to aid in understanding the loss of cyanogens during cassava processing. About 54 percent of the cyanogens in raw cassava were lost after boiling peeled cassava, but a substantial proportion remained in the water (Table 1). After slicing the boiled cassava and steeping the slices, a substantial proportion of the cyanogens was again lost.

TABLE 1 Cyanide Content of Cassava During Processing Into Ijapu

Cyanide Content (ppm)
Material analyzed	pH	Total	Free	HCN	Bound	Cyano
Unprocessed peeled cassava	6.3	76.1+15.3	5.5+2.2	2.9+0.4	70.6	2.6
Boiled cassava	6.0	35.1+8.7	2.4+1.2	2.0+0.7	32.7	0.4
		(53.9)			(53.7)
Boil water	6.2	12.8+2.1	1.1+0.2	0.5+0.2	11.6	0.5
"Ijapu"
(after 24 hours steeping)	ND	11.8+1.4	4.2+1.4	3.1+1.4	7.6
		(84.5)			(89.2)
Steep water	4.0	16.2+3.9	5.7+3.4	5.6+3.4 7.8	0.1

Note: Cyano, Cyanohydrin; numbers in parenthesis, percent loss; ND, not determined

Much of the loss could, however, be accounted for in the steep water, and the proportion lost depended on the duration of steeping and the cassava: water ratio (Tables I and 2).

ROLE OF FERMENTATION

The reduction of >99 percent in the linamarin content of grated cassava within 6 hours of adding water, with little or no change in the pH of the mash, would imply that fermentation had nothing to do with the detoxication. Linamarin breakdown is essentially a hydrolytic process catalyzed by the endogenous enzyme linamarase (1,2). The results of the present study indicate that the addition of water aids in the hydrolytic process. Apparently not all of the water normally in raw cassava tuber is available for hydrolysis.

During the boiling of cassava for processing into ijapu, linamarase would be inactivated. Yet a substantial proportion of the linamarin in cassava was still lost, appearing to a large extent in the water used for boiling and for steeping (Tables 1 and 2). This would suggest that leaching could be an important factor in cyanide loss during cassava processing. This would be true not only during the boiling and steeping of cassava for ijapu production but also during the dewatering of grated cassava for gari production.

CONCLUSION

Cassava detoxication during processing is essentially an enzymic hydrolysis of cyanogens in cassava (1,2,8). Fermentation has little role

TABLE 2 Effect of Sliced Boiled Cassava: Steep Water Ratio on Cassava Detoxification

Ratio

(w/v)	Material analyzed	pH	Total	Free	HCN	Bound	Cyano
1:1	Unboiled cassava	6.4	90.2	4.5	3.0	85.7	1.5
	Boiled cassava(a)	ND	51.7	1.5	0.5	50.3	1.0
		(42.7)				(41.3)
	Boil water	6.1	20.2	0.5	0.3	19.7	0.2
	Sliced, boiled	ND	5.6	1.1	0.6	4.5	0.6
	cassava(b)	(93.8)				(94.7)
	Steep water	4.3	25.5	5.7	4.5	19.7	1.3
1:2	Sliced boiled	ND	5.7	1.7	0.6	4.0	1.2
	cassava(b)	(93.7)				(95.3)
	Steep water	4.3	21.3	1.9	1.4	19.4	0.5
1:3	Sliced, boiled	ND	4.2	1.6	0.5	2.6	1.0
	cassava(b)	(95.3)				(97.0)
	Steep water	4.2	11.5	2.1	1.1	9.4	1.0

(a) Prior to slicing and soaking in water.

(b) After 24 hours soaking; other notes as in Table 1.

in this process and may even be antagonistic to it (10). Leaching is another important process for cyanogen reduction during cassava processing. Although fermentation does not aid in cassava detoxication during processing, it is important in flavor development (11,12) and preservation of the product.

REFERENCES

1. Conn, E.E.1969. Cyanogenic glucosides. Journal of Agricultural and Food Chemistry 17:519-526.

2. Nartey, F. 1978. Cassava: Cyanogenesis, Ultrastructure and Seed Germination. Copenhagen, Denmark: Munksgaard.

3. Cooke, R. D., and E. N. Maduagwu. 1978. The effects of simple processing on the cyanide content of cassava chips. Journal of Food Technology 13:299-306.

4. Oke, O. L. 1983. Processing and detoxication of cassava. Pp. 329-336 in: Proceedings: 6th Symposium of the International Society for Tropical Root Crops, Lima.

5. Sokari, T. G., P. S. Karibo, and L. F. F. Manuel, 1991. Substitution of boiling for steeping in ogi production. Discovery and Innovation (In press).

6. Manuel, L. F. F. 1990. Microbiological, Nutritional and Sensory Evaluation of the Effects of Cowpea Fortification of Ogi. M.Phil. thesis, Rivers State University of Science and Technology, Port Harcourt, Nigeria.

7. Sokari, T. G., P. S. Karibo, and C. K. Wachukwu. 1991. Reevaluation of the role of fermentation in cassava detoxification during processing into foods. Proceedings: Workshop on Traditional African Foods, Dar-es-Salaam, Tanzania (In press).

8. Cooke, R. D. 1978. An enzymatic assay for the total cyanide content of cassava. Journal of the Science of Food and Agriculture 29:345-352.

9. Cooke, R. D. 1979. Enzymatic assay for determining the cyanide content of cassava and cassava products. Cassava Information Center, Centro Internacional de Agricultura Tropical, Cali, Colombia, 05EC-6, 14 pp.

10. Maduagwu, E. N. 1983. Differential effects on the cyanogenic glucoside content of fermenting cassava activities. Toxicology Letters 15:355-359.

11. Vasconcelos, A. T., D. R. Twiddy, A. Westby, and P. J. A. Reilly. 1990. Detoxification of cassava during gari preparation. International Journal of Food Science and Technology 25:198-203.

12. Dougan, J., J. M. Robinson, S. Sumar, G. E. Howard, and D. G. Coursey. 1983. Some flavoring constituents of cassava and of processed cassava products. Journal of the Science of Food and Agriculture 34:874-884.

14 Solid-State Fermentation of Manioc to Increase Protein Content

Nguyen Ngoc Thao and Nguyen Hoai Huong

Manioc (cassava) is grown extensively in Vietnam and other tropical countries for its high yields in infertile soil. Although manioc is high in carbohydrates, its use is limited by its low protein content (1 to 4 percent). Manioc has been used at levels of 10 to 15 percent in poultry feed and 35 to 50 percent in pig feed. Powdered dried fish debris (gills, scale, tail, etc., from the fish processing industry), oil cake (from coconut or peanut oil production), or soybean flour have been used to raise protein levels in such feeds, but these products raise the price of feed significantly.

To upgrade the protein content in manioc, yeast cells or fungi can be inoculated in a manioc-containing medium along with nutrients containing nitrogen, phosphorus, and potassium. The use of mycelial fungi has the following advantages:

· The protein content in fermented product can increase to 30 percent.

· Fungal protein can be substituted completely for animal protein.

· The product has a low nucleic acid content.

· The product contains a favorable spectrum of amino acids.

Solid-state fermentation and liquid-state fermentation are two methods used for cultivation of fungus. Liquid-state fermentation processes are well developed in industrialized countries but are not suitable for rural farms in developing countries. Solid-state fermentation is a simple process that does not require modern equipment, power supply, or sterile conditions. In addition, the capital investment is low, permitting countryside operation and the use of available manual labor.

Many studies of solid-state fermentation of manioc have been conducted. The cultivation of Aspergillus niger in a manioc medium at 35° to 40°C for 30 hours has resulted in protein content increases of 5 to 18 percent; carbohydrate content decreased from 65 to 28 percent (1). The protein from this fermentation can be competitive with soybean protein.

In addition to A. niger, other fungi such as A. awamori, A. hennebegii, A. fugamitus, Rhizopus chinensis, and Sephalo sporium lichlorniae can be grown in acid medium at high temperatures. The protein content of the fermented product can reach 48 percent.

In Vietnam, A. niger and A. hennebegii were cultivated on a maltolized-manioc medium or a mixture of manioc and rice flour. This research comes from the demand of the husbandry industry and is designed to develop a fermentation process for on-farm use.

MATERIALS

Dried manioc pieces were ground to the size of 5 to 10 mm. Spores of A. niger were cultivated by surface fermentation on a medium containing rice hulls, rice bran, or manioc flour as carbohydrate, and urea (2 percent), ammonium sulfate (8 percent), and potassium phosphate (4 percent) at pH 4.5. Spores were collected after 7 days of cultivation.

METHODS

After a defined period of fermentation, the product was dried at 65° to 70°C, ground, and analyzed. The moisture content was determined by drying at 105°C to constant weight. The protein content was determined by precipitating with a solution of CaSO4 (6 percent) and NaOH (1.25 percent); the precipitate was analyzed by the Kjeldahl method. The starch content was determined by hydrolyzing the preparation with HCl and using the Bertrand method. The reduced glucose content was determined by the Bertrand method.

Table I shows that A. niger could not grow in medium containing urea as the only nitrogen at a concentration of 4.5 percent because of the resultant alkalinity. With (NH4)2SO4 as the N source, the pH was maintained at 4 to 5 during the fermentation. The maximum protein content was attained in medium containing urea (4 percent) and ammonium sulfate (5.8 percent). The content of protein can reach 17 percent in comparison with the one cultivated in only urea-containing medium. However, 1.55 percent N protein was achieved in the culture medium containing 3.1 percent N with the transformation efficiency of 49 percent.

TABLE 1 Effect of Nitrogen Sources on Protein Formation

N Protein

Percent Percent N Protein

N in Culture in Fermented in Fermented

N°	Source	Medium	Product	Preparation
1 Urea	2.0	0.93	5.2	0.83
	3.3	1.54	8.0	1.28
	4.0	1.86	8.2	1.3
	4.5	2.10	no growth	--
	5.0	2.33	no growth	--
2 Ammonium sulfate
(AS)	7.4	1.54	4.4	0.70
Urea 2% + AS 10%		3.05	7.2	152
Urea 3.3% + AS 4%		3.10	9.36	1 49
Urea 4.0% + AS 5.8%		3.1	9.7	1.55
Urea 4.5% + AS 7.4%		3.1	no growth

The transformation efficiency was 70 percent in medium containing only urea (4 percent).

The P and K elements (Table 2) were added to the medium containing urea 3.3 percent and ammonium sulfate 4.4 percent (1-5) or urea 3.3 percent (6-9), respectively. The results suggested that the P and K sources had no clear effect on protein formation.

In Table 3, the effect of humidity on the protein content is shown. Table 4 illustrates the effect of sterilizing conditions on the yield of protein. Table 5 shows the effect of the amount of inoculum culture on protein synthesis.

RESULTS

Manioc flour cannot be used as a carbohydrate source in the culture medium because it agglomerates and excludes air necessary for the growth of the fungal mycelium.

Manioc pieces of 0.5 to 1.0 centimeters are best for this solid fermentation method. The protein content of fermented preparation decreased 50 percent when using manioc pieces that were 1.0 to 2.0 centimeters in size.

The analysis of a fermented preparation after 2 days of fermentation, drying at 65° to 70°C, and grinding is shown in Table 6.

TABLE 2 Effects of Nutrients on Biosynthesis of Protein (The P and K elements were from chemical fertilizers)

Chemical Fertilizer	P Percent + K Percent	Protein Percent
1	3.3 + 1.0	10.21
2	2.3 + 0.5	11.34
3	2.3 + 1.5	10.41
4	4.3 + 0.5	11.46
5	4.3 + 0.5	9.3
6	3.3 + 1.0	10.0
7	2.3 + 0.5	10.0
8	1.3 + 0.5	11.25
9	0.3 + 0.5	9.62

TABLE 3 Effect of Initial Humidity on Protein Content

	Humidity of	Protein
N°	Culture Medium(a)	Percent	Notes
1	45	7.9	The change of humidity from
2	50	9.36	60 to 70 percent occurred
3	55	9.64	depending on the atmospheric
4	60	11.08	temperature and humidity.
5	65	11.23
6	70	11.37
7	75	Poor growth

a The medium for this experiment contained urea (4 percent), P (1.3 percent), and K (0.5 percent).

TABLE 4 Effect of Sterilization Conditions on Protein Production

, ,

N°	Temperature °C	Time Minutes	Percent Protein	Notes
1	100	45	8.6	The culture medium
2	100	90	7.35	can be sterilized
3	120	30	7.6	at 100°C in
4	120	45	7.50	45 minutes.
5	120	60	6.30

 

TABLE 5 Effect of the Amount of Inoculum Culture on Protein Synthesis

	Percent of Inoculum	Percent	Notes
N°	Culture	Protein
1	0.5	6.0	The maximum percent of
2	1.0	8.6	protein was achieved in 2 percent
3	1.5	10.0	of inoculum culture. There was
4	2.0	12.5	formation of black spore in
5	3.0	12.0	fermented preparation when using more than 2 percent of inoculum culture.

TABLE 6 Product Analysis

			Fermented
		Manioc Pieces	Preparation,
N°	Index	Culture-Medium, Percent	Percent
1	Protein	1.5-2.0	10.0-13.3
	Starch	33	11
	Reduce sugar	4.4	8.11
	Total sugar	8.4	13.00

CONCLUSION

This solid-state fermentation method can be used to upgrade by six to seven times the protein content in manioc pieces. The resulting fermented product contains 10 to 13 percent protein, which is suitable for use as a feed additive.

REFERENCE

1. Raimbault, M.J. 1985. Fermentation Technology 63(4):395-399.

15 Leaf and Seed Fermentations of Western Sudan

David B. Harper and M. A. Collins

Kawal, sigda, and furundu are fermented foodstuffs indigenous to the Kordofan and Darfur provinces of Western Sudan. All are produced by solid state fermentation of readily available plant materials of little or no economic value which, though unpalatable in its natural state (and indeed toxic in the case of kawal), contain protein rich in sulphur amino acids. In each case, fermentation yields a product that is not only organoleptically acceptable but also sufficiently highly regarded nutritionally by the local people to be employed as a meat substitute. As the sun-dried food can be stored indefinitely without deterioration, these fermentation processes represent a food preservation technique particularly well suited to the climate and conditions of this part of Africa. Biochemical and microbiological aspects of these fermentations and their nutritional implications have been investigated by Dirar (1), Dirar et al.(2), and Elfaki et al.(3).

PREPARATION

Kawal is prepared from the fresh leaves of a wild and reputedly toxic legume, Cassia obtusifolia, which are pounded to a paste and packed into an earthenware zeer buried to the neck in the ground in a shaded location. A layer of green sorghum leaves is placed on the surface of the paste and the zeer fitted with a lid that is sealed with mud. At intervals of 3 days the vessel is opened, the sorghum leaves removed, and the paste remixed thoroughly by hand. The repacked paste is covered with fresh sorghum leaves and the zeer resealed. After 11 to 15 days, the strongly smelling black mass is removed, molded into small balls, and dried in the sun for 5 days. The dried kawal is usually consumed in a stew with onions, okra, or other local vegetables.

The seedcake remaining after oil extraction from Sesame indicum seed is the raw material for the sigda fermentation. The bitter, indigestible seedcake made from nondecorticated seed is often used only as animal feed. In the traditional sigda process the seedcake is ground to a paste with warm water. Kambo, a local form of potash from the dried leachate of the ash of the central stems of the sorghum seed head, is frequently, but not invariably, added (3 to 20 g/kg). The mixture is packed in an earthenware vessel sealed with a cotton cloth and a close-fitting lid to minimize access of air. The fermentation lasts 3 to 7 days at=30 C° with occasional remixing, addition of water if necessary, and resealing of the container, after which the product is molded into small balls and sun-dried. Like kawal, sigda is usually consumed in a vegetable stew. A similar fermented food, furundu, is prepared from the crushed seeds of karkade (Hibiscus sabdariffa) by a process almost identical to that employed for sigda.

MICROBIOLOGY

The microflora of C. obtusifolia leaves (the substrate for the kawal fermentation) was dominated by four bacterial species, Bacillus subtilis, Lactobacillus plantarum, Propionibacterum sp., and Staphylococcus sciuri, and two yeasts, Candida krusei and Saccharomyces sp. Although the relative proportions of these organisms changed, all persisted in detectable numbers throughout fermentation. The principal species present during fermentation were B. subtilis and Propionibacterium sp., the other organisms comprising a comparatively small proportion of the population. No marked interspecial successional pattern occurred during fermentation.

The microflora of unfermented sesame seedcake was dominated by two bacterial species, Pediococcus sp. and Streptococcus sp., and two yeasts, Saccharomyces sp. and Candida sp. Pediococcus sp. was eliminated after the second day of fermentation, and the occurrence of the two yeasts was confined to the first half of the fermentation period. However, the homofermentative lactic acid bacterium Streptococcus sp. dominated the microflora throughout most of the fermentation. Additionally, the yeasts Debaryomyces sp. and Torulopsis sp. appeared in low numbers late in fermentation.

No detailed examination of the microflora during fermentation of furundu has been attempted, but the principal organism present in the final product was identified as a Bacillus sp.

PROTEIN CONTENT AND QUALITY

The crude protein content decreased only slightly, if at all, during fermentation of each substrate, indicating little loss of nitrogen during the process (Table 1). It is clear that the high sulphur amino acid content of all the fermentation substrates is largely retained in the fermented products, which compare favorably with the FAO reference protein in this respect (Table 2). The branched chain amino acids valine, leucine, and isoleucine also tend to be at a higher level in the protein of sigda and furundu than in the protein of their respective substrates. The other noteworthy feature is the markedly enhanced concentration of alanine in sigda and, to a lesser extent, in furundu, compared with the unfermented substrate. This increase is probably attributable to the transamination of pyruvate formed by oxidation of the lactic acid produced in the fermentation. Significantly, alanine concentration did not rise during the kawal fermentation where lactic acid production is negligible.

The overall protein quality of each of the fermented foods is determined by the content of lysine, which is limiting in the raw material for both the sigda and furundu fermentations and does not increase appreciably during fermentation. Nevertheless, the proteins of kawal and furundu, with chemical scores of 73 and 80, are of surprisingly good quality, whereas that of sigda, with a chemical score of 33, is no poorer nutritionally than the protein of the local staple cereal, sorghum.

MINERAL, CRUDE FIBRE, AND OIL CONTENT

Ash content of all fermented foods showed a substantial increase on that of the unfermented substrate, which, in part, reflects the mineral contribution made by clay scraped from the interior of the fermentation vessel during preparation but also, in the case of sigda and particularly furundu, the liberal addition of kambo (Table 1). The latter consists largely of potassium bicarbonate with smaller quantities of potassium chloride, silicate, and sulphate. C. obtusifolia leaves display an unusually high calcium content, which is believed to be critical in determining the course of fermentation (see below). Oil and crude fibre contents of the fermented foods was not significantly different from that of the unfermented substrates, suggesting that participation of these fractions in the fermentation process is unlikely.

CHANGES DURING FERMENTATION

The dominant role of lactic acid and the marked decrease noted in pH during the sigda fermentation contrast strongly with the high concentrations of volatile fatty acids (VFA) and minimal pH change observed in the kawal fermentation (Table 3 and Figure 1).

TABLE 1 Composition of Field Collected Kawal, Sigda, and Furundu Compared With That of Their Fermentation Substrates on a Dry Weight Basis

			Crude	Crude
	Ash	protein	Oil	fibre	KNa	Ca	Mg	P	S	Fe	Zn	Mn	Cu
	%	%	%	%	%%	%	%	%	%	mg kg-1,	mg kg-1	mg kg-1	mg kg-1
Leaves of C.	12.6	24.3	2.5	13.5	ND(a)	ND(a)3.85	0.30	0.26	ND	534	32	75	ND
obtusifolia
Kawa/	19.6	26.2	3.8	12.1	2.5ND(a)	4.13	0.42	0.28	0.52	82	84	112	11
Sesame	14.0	45.6	14.4	7.4	1.04<0.01	1.87	0.66	1.12	0.74	708117	68	38
seedcake(b)
Sigda	18.2	43.8	16.9	8.2	1.830.67	2.25	0.66	1.11	0.75	509	127	83	34
Karkade	6.2	32.6	21.1	25.1	1.27<0.01	0.31	0.43	0.62	0.36	313	90	118	18
seed
Furundu	22.8	26.5	23.3	26.5	5.650.08	0.58	0.69	1.08	0.63	347	116	122	21

(a) ND Not detemmined

(b)Assara extracted

TABLE 2 Amino Acid Composition of Kawal, Sigda, and Furundu and Their Fermentation Substrates Compared With the FAO Reference Protein

Amino add concentration (g 16 g^-1 N)

										NH2	NH2
	Asp	Thr	Ser	Glu	Pro	Gly	Ala	Val	Cys	Met	ILeu	Leu	Tyr	Phe	His	Lys	Arg	Orn	But	But
Leaves of c	12.1	6.2	4.6	13.6	7.7	6.7	7.5	7.5	1.4	2.1	6.0	10.4	5.3	6.8	3.3	7.7	7.2	<0.1	<0.1	2.6
obtusifolia
Kawal	7.7	3.3	2.8	8.2	4.2	5.0	6.8	6.4	1.2	1.5	5.1	8.3	3,5	5.4	2.0	4.0	4.0	0.1	0.7	4.1
Sesame	7.8	3.1	3.7	20.9	3.9	5.4	4.7	6.3	1.6	2.4	3.2	6.6	3.2	4.2	2.4	2.0	12.8	<0.1	<0.1	<0.1
seedcake
Sigda	7.7	2.6	3.1	20.1	4.2	6.0	9.9	6.0	2.1	2.5	4.6	8.0	2.8	4.5	2.0	1.9	10.4	<0.1	1.2	4.1
Karkade seed	11.0	3.2	4.8	24.3	4.1	5.4	4.4	4.5	2.2	3.1	3.4	6.8	3.2	4.8	2.4	4.2	13.0	<0.1	<0.1	<0.1
Furundu	10.4	3.5	3.5	20.2	4.5	5,9	6.2	5.3	1.9	2.63.6	7.1	2.6	4.3	2.0	4.4	7.7	0.9	<0.1	0.2
FAO reference			4.0						5.0	3.5	4.0	7.0	6.0			5.5
protein

TABLE 3 Lactic Acid and Volatile Fatty Acid Content of Field Collected Kawal, Sigda, and Furundu (Mean and Range in g 100 g^-1 Dry Matter)

Acid	Kawal	Sigda	Furundu
Lactic	0.21	3.07	0.50
	(0.03-0.51)	(2.85 3.35)	(0.03-1.67)
Acetic	5.08	1.10 1.59
	(2.12-6.75)	(1.00-1.19)	(1.22-2.05)
Propionic	0.90	0.04 0.09
	(0.51-1.59)	(0.03-0.05)	(0.02-0.25)
Isobutyric	0.24	<0.01	<0.01
	(0.04-0.38)
n-Butyric	2.94	0.08	0.24
	(1.18-4.73)	(0.02-0.15)	(0,05-0.73)
Isovaleric	0.22	0.02	0.17
	(0.06-0.60)	(0.01-0.03)	(0.02-0.35)
n-Valeric	0.18	<0.01	0.01
	(0.01-0.61)	(0.01-0.02)
Total VFA	9.56	1.24	2.17
	(4.70-12.1)	(1.12-1.38)	(1.65-2.63)

Thus, by the eleventh day of the latter fermentation, VFA - mainly n-butyric (8 percent), acetic (5 percent) and n-propionic (9 percent) - comprised 15 percent of the fermentation mixture. However, the pH had not changed by more than 0.5 unit from the initial value. On the other hand, by the fifth day of the sigda fermentation, when a total acid concentration of 6 percent had been attained, the pH of the fermentation mixture had fallen to 4.0 from an initial value of about 6.0. This difference in the course of fermentation is almost certainly attributable to the stronger buffering capacity of the substrate of the kawal fermentation, C. obtusifolia leaves, which possess approximately double the calcium content of sesame seedcake. Conditions in kawal do not, therefore, favor the selection of acidoduric lactic acid bacteria.

In addition to these bacteria, the two yeast species present in unfermented sesame seedcake proliferated during the initial period of fermentation. Concomitantly, starch levels were observed to fall rapidly from 2 percent in the unfermented substrate to zero after the first two days of fermentation. As the only amylolytic organisms present, the yeasts presumably were responsible for degradation of starch, rendering it available to the lactic acid bacteria. The poorly fermentative yeasts Torulopsis sp. and Debaryomyces sp. isolated in the final stages of the fermentation can utilize lactic acid aerobically and may cause the decline in concentration of the compound during this period. The addition of kambo did not appear to have any significant effect on the course of the sigda fermentation, and it was concluded that this supplementation was probably practiced mainly on organoleptic grounds.

FIGURE 1 Changes in pH, VFA, lactic acid, ammonia-N concentrations during fermentation of (a) kawal and (b) sigda.

--, pH; -·-,VFA concentration; -O-, Am, lactic acid concentration; Am, -D- ammonia-N concentration

The VFA, primarily n-butyric and acetic acids, which are the principal products of the kawal fermentation, are characteristic of clostridial fermentation of plant material as is the accumulation of ammonia nitrogen. However, all attempts to isolate clostridial species from the fermentation mixture were unsuccessful, indicating that the microbial origin of VFA must be sought elsewhere. The formation of acetic acid can probably be ascribed to the heterofermentative B. subtilis, the co-dominant microorganism, whereas propionicacid is probably an end product of anaerobic fermentation by propio'Zi~acte ~ium sp. for which lactate is a preferred substrate. Utilization of lactate in this way could explain the low level of lactate in Karl, despite the substantial population of Lactoloacill~`s plantar``m. The microbial pathway leading to formation of n-butyric acid is difficult to define, although its production may be characteristic of fermentation by this type of mixed culture as a whole rather than that by any single microorganism. n-Propanol (2.3 percent), n-butanol (01 percent), and ethanol (0.1 percent) were also detected in Karl toward the end of fermentation, though all were lost from the product during the drying phase. Formation of such alcohols is probably due to anaerobic fermentation of carbohydrate by the yeast species present.

The identification of Bacillus sp. as the principal microorganism in furundu when considered in the context of a final pH of 6.2 and the presence of both VFA and lactic acid in the fermentation mixture suggest that the furundu fermentation may be intermediate in character between those of sigda and kawal. Further investigation of thefuru,~du fermentation would be most instructive in this respect.

CONCLUSIONS

The sigda and furundu fermentations appear quite unlike the traditional oilseed fermentations practiced in Nigeria and elsewhere in West Africa where foods such as ogili and ogiri are fermented from castor oil seed (Ricinus communis) and melon seed (Citrullus vulgaris). There are even variations of the fermentation which use sesame seed and karkade seed known as ogiri-sara and red sorrel, respectively. During these West African fermentations, the pH increases to over 9.0 and ammonia production is frequently observed in the later stages. The fermentations are dominated by Bacillus s p., frequently Bacillus s`~btilis, an organism associated with spoilt sigda in Sudan. The principal reason for the difference would appear to be in the preparation of the seeds prior to fermentation, which in West Africa involves boiling for several hours in water until soft. Such pretreatment may alter the course of fermentation by two mechanisms - first, by rendering protein and polysaccharide more available for degradative attack by microorganisms, and second, by effectively eliminating much of the heat-sensitive indigenous microflora. The removal of amylolytic yeasts may well favor the selection of amylase-producing bacteria such as Bacillus sp. rather than lactic acid bacteria incapable of utilizing starch.

The three fermentations studied appear to afford a route by which unpalatable plant material or oilseed cake of little economic value can be converted into acceptable meaty-tasting food that is particularly rich in sulphur amino acids, which tend to be deficient in diets where access to meat or fish is limited. Phytic acid present in seeds can frequently hinder absorption of minerals in the gastrointestinal tract. As fermentation of plant products has been shown to reduce physic acid levels substantially, it is likely that the bioavailability of minerals in both sesame and karkade seed is increased in sigda and furundu.

REFERENCES

1. Dirar, H. A. 1984. Kawal meat substitute from fermented Cassia obtusifolia leaves. Economic Botany 38:342-349.

2. Dirar, H. A., D. B. Harper, and M.A. Collins. 1985. Biochemical and microbiological studies on kawal, a meat substitute derived by fermentation of Cassia obtusifolia leaves. Journal of the Science of Food and Agriculture 36:881-892.

3. Elfaki, A. E., H. A. Dirar, M. A. Collins, and D. B. Harper. 1991. Biochemical and microbiological investigations of sigda - a Sudanese fermented food derived from sesame oilseed cake. Journal of the Science of Food and Agriculture 57.

16 Continuous Production of Soy Sauce in a Bioreactor

Takashi Hamada, Yaichi Fukushima, and Hiroshi Motai

Soy sauce is a traditional all-purpose seasoning with a salty taste and sharp Havor. In the conventional method of brewing soy sauce (Figure 1), cooked soybeans and roasted wheat are mixed with spores of Aspergillus species and fermented in solid culture for 2 days to produce koji. The koji is then mixed with brine to make moromi, the mash that ferments to produce soy sauce. Over time the soybeans and wheat are hydrolyzed by enzymes such as proteinases, peptidases, and amylases. During the first stage of moromi fermentation, Pediococcus halophilus grows and produces lactic acid, which lowers the pH. Accompanying the decrease in pH, vigorous alcohol fermentation by Zygosaccharomy-ces rouxii occurs. As a result, 2 to 3 percent ethanol and many kinds of aroma components are produced by this yeast. At the same time, phenolic compounds such as 4-ethylguaiacol (4EG) and 4-ethylphenol, which add characteristic aroma to soy sauce, are produced by other types of yeasts such as Candida versatilis and Candida etchellsii.

It takes over 6 months for the entire fermentation and aging of the morommi mash. Therefore, shortening this period is important and new processes for soy sauce brewing are desirable. This paper describes the continuous production of soy sauce in a bioreactor system, which consists of reactors containing immobilized glutaminase and immobilized cells of P. halophilus, Z. rouxii, and C. versatilis.

MANUFACTURING PROCESSES

The processes for soy sauce production using the conventional and bioreactor methods are shown in Figure 1. The bioreactor method differs from the conventional one in the following ways: (a) proteases from continuous submerged culture are used (1), (b) fermentation is carried out in the liquid state, and (c) the fermentation period is considerably shorter. It takes several months for the conventional fermentation but only about 2 days for the bioreactor method.

In the bioreactor method, raw liquid was successively passed through, first, a glutaminase reactor to increase glutamic acid; second, a P. halophilus reactor to carry out lactic acid fermentation; and, third, a Z. rouxii reactor to carry out alcohol fermentation and a C. versatilis reactor to produce phenolic compounds such as 4-ethylguaiacol. Two reactors containing immobilized yeast cells were set in parallel, and the flow rate of the feed solution to the Z. rouxii and C. versatilis reactors was set in a ratio of 10 to 1. Carrier, packed gel volume, and operating conditions such as residence time, temperature, and aeration in each reactor are shown in Table 1.

FIGURE 1 Manufacturing processes for soy sauce by conventional and bioreactor methods.

TABLE 1 Conditions of Fermentation in Each Reactor

		Column	Packed Gel
		Volume	Volume	Residence
Reactor Carrier		(L)	(L)	Time,hours	Temperature, °C	Aeration,wm
Glutaminase	Chitopearl	1.8	0.6	0.7	40	--
P. halophilus	AS	7.5	5.0	61	27	--
Z. rouxli	Al	27.0	8.0	25.5	27	0.005
C. versatilis	Al	1.0	0.2	10.7	27	0.08

AS, Alginate-colloidal silica. Al, Alginpte.

CONTINUOUS FERMENTATION

A profile of continuous fermentation by immobilized cells of P. halophilus, Z. round, and C. r~ersatilis is shown in Figure 2. The fermentation continued for over 100 days without any microbial contamination. A consistent increased level of glutamic acid (in the range of 0.3 to 0.4 percent) was found in the effluent from glutaminase reactor, with a residence time of 0.7 hours. Lactic acid was produced by immobilized cells of P. halophilus in quantities of 0.7 to 1.0 percent at a residence time of about 6 hours, and consequently the pH declined to 4.9 to 5.0, similar to that of conventionally brewed soy sauce. Ethanol was produced constantly by immobilized cells of Z. rotlxti in quantities of 2.5 to 2.7 percent at a residence time of about 26 hours. This is the standard ethanol content in soy sauce. About 10 ppm (parts per million) of 4-ethylguaiacol was produced by immobilized cells of C. versatilis at a residence time of about 10 hours, and the final 4ethylguaiacol content after mixing the two fermented liquids from the reactors of Z. rouxii and C. versatilis was about 1 ppm, which is the optimum concentration in conventional soy sauce. The total residence time for lactic acid and alcohol fermentation was about 30 hours in this system. This was considerably shorter than the conventional fermentation period of 3 to 4 months required to produce the same amounts of lactic acid and ethanol.

High numbers of viable cells were present in the gel and liquid in each reactor. The number was 10- to 100-fold higher in moromi mash. The shortening of the fermentation period in the bioreactor method is possibly due to the high density of immobilized cells in the gel and free cells in the liquid.

The main chemical components of the fermented liquid from the bioreactors were examined, including lactic acid, glucose, ethanol, and nitrogenous compounds.

FIGURE 2 Profile of continuous fermentation of soy sauce by a bioreactor system. (), lactic acid; (a), glutamic acid (values indicate the increase in the amount of glutamic acid); (v) pH; (•), ethanol; (D), glucose; 4-KG after passing through the C. versatilis reactor; 4-KG in the final product.

PROPERTIES

The organic acids and aroma components in the bioreactor soy sauce were examined. The proportions of organic acids except citric acid were not much different between the bioreactor soy sauce and the conventional one, although the former was a little lower in acetic acid and succinic acid. It appears that the high residual content of citric acid in the bioreactor soy sauce arises from the inability of P. halophilus to utilize citric acid. Aroma components present in both the bioreactor and conventional soy sauces were not qualitatively different. However, the former was higher in isoamyl alcohol and acetoin and lower in isobutyl alcohol, ethyl lactate, 4-hydroxy-2(orS)-ethyl-5(or2)-methyl3(2H)-furanone, and 4-hydroxy-5-methyl-3(2H)-furanone.

To evaluate the aesthetic qualities of the bioreactor-produced soy sauce, sensory tests were carried out. For example' the intensity of the alcoholic, fresh, sweet, acid, and sharp odors as well as the special lZiga (baking aroma) and bushoshu (foul fermented aroma) were compared between the bioreactor and conventional soy sauces. The odors are important for the quality of soy sauce. Although the bioreactor soy sauce was a little weaker in aroma and fresh odor than the conventional soy sauce, the quality of the former was generally judged to be similar to that of the latter.

The total time required for the production of soy sauce by the bioreactor system, including enzymatic hydrolysis of the raw materials, fermentation with immobilized whole cells, and the refining process, is only about 2 weeks (2). This is considerably shorter than the 6 months with the conventional method of soy sauce brewing consisting of koji making, fermentation and aging of morons, and refining.

From these results we conclude that the quality of the bioreactor soy sauce was very similar to that of the conventional soy sauce from both chemical and sensory evaluations and that the bioreactor system is practical for the production of soy sauce.

REFERENCES

1. Fukushima, Y., H. Itoh, T. Fukase, and 1I. Motai. 1989. Applied Microbiology and Biotechnology 30:604-608.

2. Hamada, T., M. Sugishita, Y. Fukushima, and H. Motai. 1991. Process Biochemistry 26:39-45.