Edited by Foxit Reader Copyright(C) by Foxit Software Company,2005-2007 For Evaluation Only.

International Journal of Food Microbiology 75 (2002) 197 212 www.elsevier.com/locate/ijfoodmicro

Appropriate starter culture technologies for small-scale fermentation in developing countries

W.H. Holzapfel*
IBM/IHT, BFE, Haid-und-Neu-Strasse 9, D-76131 Karlsruhe, Germany Accepted 28 December 2000

Abstract Modern food biotechnology has moved a long way since ancient times of empirical food fermentations. Preservation and safeguarding of food are, however, still major objectives of fermentation. In addition, other aspects, such as wholesomeness, acceptability and overall quality, have become increasingly important and valued features to consumers even in developing countries where old traditions and cultural particularities in food fermentations are generally well maintained. Due to limitations in infrastructure and existing low technologies, rural areas in most developing countries have not been able to keep abreast of global developments toward industrialisation. At the same time, fermented foods play a major role in the diet of numerous regions in Africa and Asia. In many traditional approaches, the advantages of some form of inoculation of a new batch, e.g. by back-slopping or the repeated use of the same container (e.g. a calabash) is appreciated and generally practised. Still, the benefits of small-scale starter culture application as a means of improved hygiene, safety and quality control, in support of HACCP approaches, are not yet realised in small-scale fermentation operations. Approaches and considerations for the selection of pure cultures for small-scale, low-tech applications may differ in some respects from the large-scale industrial approaches practised since 100 years. Selection criteria should take account of the substrate, technical properties of the strain, food safety requirements and quality expectations. Lack of experience in the application of starter cultures in small-scale operations and under rural conditions presents a major obstacle but also an exciting challenge to food microbiologist and technologist. Culture preservation, maintenance and distribution demand special logistic and economic considerations. Quality, safety and acceptability of traditional fermented foods may be significantly improved through the use of starter cultures selected on the basis of multifunctional considerations, also taking into account the probiotic concept and possibilities offered for improved health benefits. D 2002 FAO-AGS. Published by Elsevier Science B.V. All rights reserved.
Keywords: Food preservation; Traditional fermentation; Functional properties; Lactic acid bacteria (LAB); Food safety; Selection criteria

1. Introduction 1.1. Food fermentation Together with drying and salting, fermentation is one of the oldest methods of food preservation. Its importance in modern-day life is underlined by the wide spectrum of fermented foods marketed both in

Tel.: +49-721-6625-450; fax: +49-721-6625-453. E-mail address: wilhelm.holzapfel@bfe.uni-karlsruhe.de (W.H. Holzapfel).

0168-1605/02/$ - see front matter D 2002 FAO-AGS. Published by Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 1 6 0 5 ( 0 1 ) 0 0 7 0 7 - 3

Edited by Foxit Reader Copyright(C) by Foxit Software Company,2005-2007 For Evaluation Only.
198 W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212

developing and industrialised countries, not only for the benefit of preservation and safety, but also for their highly appreciated sensory attributes. Fermented foods are treasured as major dietary constituents in numerous developing countries primarily because of their keeping quality under ambient conditions, and also for their safety and traditional acceptability. As a technology, food fermentation dates back at least 6000 years, and probably originated from microbial interactions of an acceptable nature. Fermentation has enabled our ancestors in temperate and cooler regions to survive winter season and those in the tropics to survive drought periods, by improving the shelf life and safety of foods. Through the ages, fermentation has had a major impact on nutritional habits and traditions, on culture and on the commercial distribution and storage of food. Traditional fermentation process still serves as a substitute where refrigeration or other means are not available for the safekeeping of food. Fermented foods can, in general, be described as palatable and wholesome foods prepared from raw or heated raw materials. They are generally appreciated for attributes such as pleasant flavour, aroma, texture and improved cooking and processing properties. Microorganisms, by virtue of their metabolic activities, contribute to the development of characteristic properties such as taste, aroma, visual appearance, texture, shelf life and safety. Enzymes indigenous to the raw materials may play a role in enhancing these characteristics (Hammes, 1990). Through trial and error, traditional skills have been developed for controlling technical parameters during fermentation processes. Experience has also shown that back-sloping, or the inoculation of raw materials with a residue from a previous batch, accelerates the initial phase of fermentation and results in the promotion of desirable changes during the fermentation process.

mentation process. Being adapted to the substrate, a typical starter facilitates improved control of a fermentation process and predictability of its products (Holzapfel, 1997). In addition, starter cultures facilitate control over the initial phase of a fermentation process. 2.2. Traditional approaches: spontaneous fermentations and back-slopping Modern starter cultures are selected either as single or multiple strains, specifically for their adaptation to a substrate or raw material. Spontaneous fermentations, i.e. processes initiated without the use of a starter inoculum, have been applied in food preservation for millennia and were elucidated through trial and error, perhaps over thousands of years. The majority of small-scale fermentations in developing countries and even some industrial processes such as sauerkraut fermentations are still conducted as spontaneous processes. Various types of starter cultures and even backslopping are widely used in fermentation processes, even in industrialised countries (Table 1). Spontaneous fermentations typically result from the competitive activities of a variety of contaminating microorganisms. Those best adapted to the food substrate and to technical control parameters, eventually dominate the process. The production of metabolites (e.g. organic acids) inhibitory to other contaminating microbes (e.g. Enterobacteriaceae) may provide an additional advantage during fermentation. Bacteria typically dominate the early stages of fermentation processes, owing to their relatively high growth rate, followed by yeasts, in substrates that are rich in fermentable sugars. In numerous traditional processes, material from a previous successful batch is added to facilitate the initiation of a new process. Through this practice of backslopping, the initial phase of the fermentation process is shortened and the risk of fermentation failure reduced. Repeated use of back-slopping results in selection of the best-adapted strains, some of, which may possess features that are desirable for use as starter cultures. 2.3. Inoculation to improve process control

2. Development of concepts toward the use of starter cultures 2.1. Definitions A starter culture may be defined as a preparation or material containing large numbers of variable microorganisms, which may be added to accelerate a fer-

Initiation of a spontaneous fermentation process takes a relatively long time (24 48 h), with high risk

Edited by Foxit Reader Copyright(C) by Foxit Software Company,2005-2007 For Evaluation Only.
W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212 Table 1 Starter cultures, typical of industrialised countries and used in various fields of food fermentation Foodstuff Sauerkraut Various vegetables Vegetable juices Soy products Sour dough Wine Dry sausage Dairy products Single-strain cultures + + + + + + + + Multiple-strain cultures + + + + Mixed-strain cultures + + Back-slopping +a + + () 199

, not used; +, applied. Source: Buckenhuskes, 1993. a Brine from a previous fermentation.

for failure. During this early phase which is associated with the lag phase of microbial growth, contaminating microorganisms on raw materials, utensils and from the environment, slowly increase in number and compete for nutrients in order to produce metabolites. This phase can be shortened by inoculation either through back-slopping or with the use of selected starter cultures. Beneficial attributes of the substrate, consumer expectations and technical requirements dictate to a large extent the nature of the starter culture to be used, i.e. single-strain vs. mixed-strain culture. 2.4. Hygiene, safety and quality considerations in support of HACCP approaches in small-scale fermentation processing Fermentation is generally considered as a safe and acceptable preservation technology for improving the hygienic quality and safety of foods. Failure of fermentation processes can, however, result in spoilage and/or the survival of pathogens, thereby creating unexpected health risks in food products, which would otherwise be considered safe. Inoculation with starter cultures does not provide an absolute guarantee against failure of fermentation processes, nor does it eliminate health hazards associated with pathogens, toxinogens, toxic components or residues. Metabolic activities of desirable fermentation microorganisms must be supported by observing the basic principles of Good Manufacturing Practice (GMP). This implies the maintenance and control of technical parameters that ensure the desired outcome of the fermentation process. Precautions should also be taken against the introduction or transfer of potential health hazards or

factors that are potentially detrimental to quality during the fermentation process. The Hazard Analysis Critical Control Point system (HACCP), presents a scientific and systematic approach for enhancing the safety of foods, from primary production to final consumption, through the identification, evaluation and control of hazards that are of significance for food safety (Amoa-Awua et al., 1998; WHO, 1995). Application of HACCP to fermentation is thoroughly covered elsewhere in this publication (Motarjemi and Asante, 2002).

3. Selection of starter cultures and their application in small-scale fermentations 3.1. Approaches and considerations for the selection of pure cultures Considerations for applying starter cultures at the household level should take into account cultural traditions, dietary habits and raw materials, which differ across regions and continents. Lactic fermented cereals produced by small-scale spontaneous solid- or semisolid-state fermentations are widely accepted and appreciated by consumers in most African countries. In Southeast Asia, on the other hand, traditional fermentations are reliant on moulds as the dominant organism, and on legume food substrates such as soya beans. From a technical viewpoint, cost/benefit ratios, logistical factors and the willingness of the small-scale processor to accept new approaches is critical in any assessment of the feasibility of introducing the use of starter cultures in small-scale fermentations. In addi-

200

W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212

tion, a minimum set of standards or quality parameters for the handling and maintenance of these cultures would need to be developed for use at the artisanal level. The introduction of starter cultures should be considered within the context of realistic prospects for: information transfer minimal technical adjustments to small-scale low-tech food fermentations, application of the HACCP system and education and training. The prospect of applying starter cultures will become attractive to the small-scale processor only if benefits, such as reduction of costs (e.g. energy), reduced fermentation times, reduced risk of spoilage (increased shelf-life), improved process control, improved sensory quality (taste, aroma, visual appearance, texture, consistency), improved safety attributes (e.g. lower risk if diarrhoea, detoxification of cassava) and reduced preparation procedures for the final product, are perceived. Some LAB and yeast strains associated with fermented foods, are capable of degrading antinutritional factors, such as phytic acid and phenolic compounds. Incorporation of these organisms into starter cultures may, therefore, to serve upgrade the nutritional value of foods. Furthermore, selected strains may enhance the general benefits of spontaneous fermentation such as improved protein digestibility and micronutrient bioavailability, and contribute more specifically to biological enrichment through the biosynthesis of vitamins and essential amino acids. In recent times, there has been a considerable focus on the inclusion of mycotoxin-degrading strains in starter cultures. The use of mycotoxin contaminated raw materials for fermentation in developing countries, poses a special challenge for the selection of strains that are capable of mycotoxin detoxification (Adegoke et al., 1994; Smith et al., 1994; Westby et al., 1997; Holzapfel et al., 1998). The probiotic properties of strains involved in food fermentations are also being studied in light of their potential contribution to the improvement of general health and well being. 3.2. Selection criteria for starter culture development Spontaneous food fermentations are neither predictable nor controllable. Pure cultures isolated from mixed populations of traditional fermented foods exhibit a diversity of metabolic activities, which vary even among strains. These include differences in

growth rate, adaptation to a particular substrate, ability to degrade antinutritive factors, antimicrobial properties, flavour and quality attributes and competitive growth behaviour in mixed cultures (Holzapfel, 1997). Single-and mixed-strain cultures must, therefore, be tested at the pilot scale, prior to their use in small-scale operations. According to Holzapfel (1997), the introduction of starter cultures in traditional small-scale fermentations should incorporate considerations for improving processing conditions and product quality through: (i) rapid accelerated metabolic activities (acidification or alcohol production); (ii) improved and more predictable fermentation processes; (iii) desirable sensory attributes; (iv) Improved safety and reduced hygienic and toxicological risks. Commercial starter cultures generally originate either from food substrates or from the processes in which they are applied. Environmental conditions, back-slopping, adaptation and the repeated use of specific utensils can contribute to the selection of microbial populations typical of a fermentation process. The selection of suitable starter strains should take into account their interactions in mixed cultures, with consideration for the behaviour of these strains under defined conditions, and within the food substrate. Other factors, which should be considered, include: (i) competitive behaviour, viability and survival; (ii) antagonism against pathogens and spoilage microbes; (iv) the rate of acid or alcohol production; (v) organoleptic changes; (vi) primary metabolites of fermentation; (vii) degradation of antinutritive factors; (vii) detoxification; (viii) probiotic features (Holzapfel, 1997). Modern approaches incorporate considerations for technical safety and health-promoting features in the selection of the most optimal strain(s) for a process. Ideally, a multifunctional strain is targeted. The rationale for such novel approaches is now discussed. 3.2.1. Technical considerations Technical aspects of starter culture development should incorporate considerations relevant to adoption of the starter to the substrate, the rate of acid production, fermentation metabolites (e.g. hetero- vs. homofermentation) and the ability of single or mixed strain cultures to produce desirable sensory qualities in the fermented product. Numerous reports indicate that Lactobacillus brevis, L. fermentum, L. plantarum, L. reuteri, Pedio-

W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212

201

coccus pentosaceus and P. acidilactici exhibit superior performance in lactic fermented cereal and vegetable products (Steinkraus, 1996, 1997; Holzapfel, 1997; Lee, 1997; Oyewole, 1997). This is quite possibly the case for root crops, while the initiation of milk fermentations is typically associated with Lactococcus lactis, followed by L. casei (paracasei) and other Lactobacillus spp. during maturation. Several LAB are associated with meat and fish fermentations. L. sakei and L. curvatus (Hammes and Hertel, 1998) have been determined to be superior starter cultures for meat fermentations. Leuconostoc mesenteroides frequently dominates the early stages of most spontaneous fermentations. Plant materials containing fermentable sugars provide suitable substrates for the yeast species Saccharomyces, Candida, Torula and Hansenula. Although the growth rate of these yeasts is lower than that of bacteria, such as L. mesenteroides, strains of Saccharomyces cerevisiae eventually dominate most spontaneous alcoholic fermentations as in the production of African opaque beers, palm wine and Asian beverages, such as rice wines and Indonesian tape. Indeed, most traditional fermentations results from the combined metabolic activities of different types of microorganisms. Hounhouigan et al. (1994, 1999) reported a stimulating effect of the yeast Candida krusei on L. fermentum and L. brevis during a mixed starter culture fermentation of the fermented maize product, mawe. 3.2.2. Antagonism This is the combined effect of different biological factors, resulting from metabolic activities of micro-

organisms and their competitive interactions. Fermentations involving yeasts (alcoholic fermentations of beer and palm wines), moulds (e.g. tempe fermentations), particular bacilli [alkaline fermentations e.g. as for dawadawa (Nigeria and Ghana), soumbala (Burkina Faso), natto (Japan) or kinema (Himalayas)] and LAB are generally recognised as safe (GRAS). Dominant microbes of these fermentations do not appear to be associated with any health risks. These beneficial microorganisms serve to some extent in safeguarding against pathogens and spoilage organisms. Lactic fermented foods in particular are considered to be safe and wholesome. On the other hand, acidification to pH values of less than 4.2 constitutes a major safety concern in fermented foods. Recent observations, however, confirm that a number of metabolites, such as acetic acid (from heterofermentative LAB), hydrogen peroxide and bacteriocins, produced during the fermentation process, exhibit antimicrobial properties which may contribute to the safety of lactic fermented foods (Table 2). Organic acids, which show strong antagonistic effects in the undissociated form at lower pH values, are particularly effective in inhibiting Gram-negative bacteria, such as pathogens. 3.2.2.1. Bacteriocins. Bacteriocins are antimicrobial substances of a proteinaceous nature that are active against closely related bacteria. They exhibit a narrow of activity but are not active against Gram-negative bacteria. Proteolytic enzymes present in food substrate are capable of inactivating bacteriocins. Bacteriocins from LAB may be classified into three structural groupings on the basis of their physico-chemical and

Table 2 Metabolic products of lactic acid bacteria which exhibit antimicrobial properties Product Organic acids Lactic acid Acetic acid Hydrogen peroxide Low-molecular metabolites Reuterin (3-OH-propionaldehyde) Diacetyl Fatty acids Bacteriocins Nisin Other Source: Holzapfel et al., 1995. Main target organisms Putrefactive and Gram-negative bacteria, some fungi Putrefactive bacteria, clostridia, some yeasts and fungi Pathogens and spoilage organisms, especially in protein-rich foods Wide spectrum of bacteria, moulds and yeasts Gram-negative bacteria Different bacteria Some LAB and Gram-positive bacteria, notably endospore-formers Gram-positive bacteria, inhibitory spectrum according to producer strain and bacteriocin type

202

W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212

antimicrobial properties (Schillinger et al., 1995; Holzapfel et al., 1995). Though relatively uncommon in fermented foods (Olasupo et al., 1994), bacteriocinogenic LAB strains are of special interest in view of their possible application in food safety assurance. Bacteriocinogenic LAB have been shown to effectively inhibit the growth of pathogens, such as Listeria monocytogenes, Staphylococcus aureus, Bacillus cereus and Clostridium dificile, even under in situ conditions (Holzapfel et al., 1995). 3.2.3. The lactic acid isomer The lactic acid isomer produced during fermentation is typically related to the LAB species from which it is produced. LAB genera, examples of which include Streptococcus, Lactococcus, Enterococcus and Carnobacterium produce > 90% of the L(+)isomer as an end product of sugar fermentation. Leuconostoc spp. and L. delbrueckii (all subspecies) on the other hand produce D( )-lactic acid. Lactic acid isomers produced by lactobacilli and pediococci are species specific. The L(+)-isomer is produced by L. casei, for example, while a racemate (DL) is produced for L. sakei, all heterofermentative lactobacilli and practically all Weissella spp. The nature of the lactic isomer is of concern, since high levels of the D( )-lactic acid isomer are not hydrolysed by LDH enzymes in humans and are, thus, capable of causing acidosis. WHO recommendations indicate a maximum daily intake of 100 mg/kg body weight of this nonphysiological lactic acid isomer (WHO, 1968). There are, however, no recommended limitations for the intake of the L(+)-lactic acid isomer. Consumption of 1 l of a fermented gruel containing 1% of DL-lactic acid, by an individual having a 50 kg body weight, could potentially result in the intake of the maximum recommended levels of this nonphysiological acid. L(+)-lactic acid producing strains should, therefore, be preferentially selected for the fermentation of beverages. 3.2.4. Antinutritive factors Antinutritive components are of particular significance in unbalanced diets, such as cereal-based diets. Cereal staples (maize, sorghum and millet) which are often admixed with legumes in order to upgrade their protein content, contain a number of antinutritive fac-

tors. Protease and amylase inhibitors, polyphenols (from millets and sorghum) and lectin-related haemagglutinin activities in legumes and tannins, adversely affect the protein and starch availability of these foods. Furthermore, the chelating properties of phytic acid may significantly reduce the bioavailability of minerals, such as calcium, iron, magnesium and zinc. These antinutritional factors, coupled with the lysine, tryptophane and methionine deficiencies in cereal proteins contribute to malnutrition in developing countries (Holzapfel, 1997). Fermentation may serve to improve the nutritional value of cereal staples through the reduction of antinutritive factors, as reported for a number of foods of plant origin (Chavan and Kadam, 1989; Lorri, 1993; Mbugua et al., 1992). 3.2.4.1. Proteinase inhibitors. Lactic fermentation has been shown to lower the levels of proteinase inhibitors in cereal porridges thereby increasing the availability of essential amino acids, such as lysine, leucine, isoleucine, methionine and even tryptophane (Kazanas and Fields, 1981; Mbugua, 1986; Nche, 1995). It is effective in reducing proteinase inhibitors (e.g. trypsin inhibitor) in legumes and tannins, and disulphide cross linkages in sorghum prolamine proteins (Hamaker et al., 1987; Khetarpaul and Chauhan, 1989). Combined lactic and yeast fermentations have also been shown to improve the protein digestibility of cereal porridges (Graham et al., 1986; Lorri, 1993; Mbugua et al., 1992) and may, therefore, serve to improve the protein quality of cereal grains. LAB strains, however, differ in their ability to degrade trypsin inhibitor under defined conditions. An approximately 50% reduction in trypsin inhibitor activity was observed in our laboratories for L. plantarum strain 91 and Leuconostoc sp. 106 (Table 3), isolated from Ghanaian fermented foods. A kinetic study of L. mesenteroides 92 activity showed that a significant decrease in trypsin inhibitor activity was affected only during the stationary phase of growth, indicating the importance of the length of the fermentation process (Holzapfel, 1997). 3.2.4.2. Phytic acid and tannins. Phytic acid and tannins are antinutritive components typical of cereal and legume foods. These compounds are of concern since they may reduce both iron and mineral bioavailability in cereal and legume-based diets. Essential steps in traditional household-level processing, such as soak-

W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212 Table 3 Degradation of trypsin inhibitor (TI) by lactic acid bacteria isolated from aflata in Ghanaa LAB isolate L. plantarum 91 L. fermentum 103 Pediococcus sp. 90 Pediococcus sp. 19 Leuconostoc sp. 106 Lactobacillus sp. 41 Laactobacillus sp. 17 Lactobacillus sp. 62 Reduction of TI (mg) 2.41 1.22 0.89 1.08 2.68 0.65 1.86 1.34 Percent reduction 48.0 24.4 17.8 21.6 53.6 13.0 37.2 26.8

203

a Investigations were conducted in a synthetic liquid medium containing 5 mg TI/ml. Incubations were conducted at 30 jC for 5 days. The TI concentration was determined according to Kakade et al. (1974) with synthetic benzoyl DL-arginine-p-nitro anilide as substrate (Holzapfel, 1997).

chyose to be relatively slowly hydrolysed. Soaking (Ogun et al., 1989) and germination (Abudu and Akinyele, 1990; Trugo et al., 1990) were determined to be important processing steps for reducing the oligosaccharide content of legumes. The production of a-galactosidase by LAB species, such as L. mesenteroides ssp. mesenteroides and ssp. dextranicum and Weissella paramesenteroides, appears to be variable (Milliere et al., 1989), whilst it appears to be a constitutive property for L. fermentum, L. brevis, L. buchneri, L. cellobiosus and L. salivarius (Mital et al., 1973) and is probably inducible in L. plantarum (ATCC 8014) (Ahrne and Molin, 1991). L. plantarum strains isolated from fermented Ghanaian maize products were able to ferment raffinose, while strains of P. acidilactici and P. pentosaceus were unable to do so (Table 4). 3.2.5. Degradation or inactivation of natural toxins The degradation or inactivation of toxins by pure cultures during fermentation has received considerable attention in recent times. Naturally occurring toxins, such as the cyanogenic glucosides (linamarin and lotaustralin) in cassava, may cause severe intoxications following the consumption of raw or unprocessed bitter cassava (Holzapfel, 1997). Detoxification during the fermentation of cassava is brought about primarily by microbial activity (Westby and Choo, 1994), although endogenous linamarinase enzymes present in cassava play a significant role in the process. Lactobacillus spp. (Amoa-Awua et al., 1996; Olasupo et al., 1997), Bacillus spp. (Ejiofor and Okafor, 1981; Essers et al., 1995; Amoa-Awua and Jakobsen, 1995) and yeasts and moulds (Hahn, 1989, Essers et al., 1995) play an important role in cassava processing. A comparison of the effects of spontaneous fermentation, back-slopping and the use of starter cultures for the reduction of cyanogenic glucosides in cassava (Kimaryo et al., 2000) revealed that all three types of fermentations contributed significantly to the detoxification of cassava, with a starter culture consisting of L. plantarum strains giving the best results. 3.2.6. Mycotoxins Mycotoxins, in particular aflatoxins and fumonisins pose a major risk in stored cereals, which are raw typical materials for traditional fermented foods in most African countries (Holzapfel, 1997). Even with strict regulations on maximum tolerable levels, control

ing, germination and lactic fermentation, may contribute to a reduction of these inhibitors. Soaking has been shown to activate endogenous phytases in most cereals and legumes. Although lactic fermentation has also been shown to reduce the phytate content of white sorghum (Svanberg and Sandberg, 1988), non-tannin containing cereals (Svanberg et al., 1993), maize (Lopez et al., 1983), pearl millet (Mahajan and Chauhan, 1987; Khetarpaul and Chauhan, 1989) and idli, a fermented cereal legume product (Reddy et al., 1986), phytic acid degrading ability is relatively rare among pure LAB cultures. Some L. plantarum strains are, however, capable of degrading phytic acid on incubation at 37 jC for 120 h (Holzapfel, 1997). Phytase activity is not detectable for Bacillus spp. associated with the fermentation of the African locust bean Parkia biglobosa which is used in the preparation of iru or dawadawa (soumbala) (Aderibigbe and Odunfa, 1990). 3.2.4.3. Oligosaccharides. Raffinose, stachyose and verbascose are oligosaccharides that typically occur in legumes and cereals, and cause flatulence, diarrhoea and indigestion in humans. These sugars possess a-Dgalactosidic bonds which are resistant to cooking and small-scale processing, but which can be hydrolysed by a-galactosidases produced by a number of moulds and by bacteria associated both with the digestive tract and with fermented foods. Extensive studies using both pure and mixed cultures have shown that neither sucrose nor raffinose is utilised by the tempe mould Rhizopus oligosporus (Sorensen and Hesseltine, 1966) although Shallenberger et al. (1967) determined sta-

204

W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212

Table 4 Raffinose fermentation by selected LAB strains isolated from fermented maize products of Ghana Species Lactobacillus plantarum Lactobacillus plantarum Lactobacillus plantarum Pediococcus pentosaceus Pediococcus acidilactici Leuconostoc mesenteriodes ssp. Mesenteroides Strain nos.a 1, 7, 8, 21, 31, 38, 43, 46, 47, 64, 65, 69 37, 39, 44, 53, 55, 70, 91 58, 60, 68 90 100 92 DSM 20343T Acid production ++ + ++

Acid production: = negative; + = weak; + + = strong. Qualitative fermentation tests were performed in MRS broth (pH 6.4) containing 1% raffinose (instead of glucose) and 0.004% chlorophenol red as indicator. Incubation was conducted at 30 jC for 3 days (Holzapfel, 1997). a Origin of strains. Nos. 1 21, 92: from fermenting maize (Aflata) during Kenkey production. Nos. 31 47: from fermenting maize/white cowpeas mixture (70:30). Nos. 53 70: from fermenting maize/red cowpeas mixture (70:30). No. 90: from Yakeyake after fermentation. No. 91: from Agbelima, after cooking. No. 100: from Aflata (for Ga-Kenkey) after fermentation.

mechanisms for mycotoxins in developing countries are inadequate, and can neither be applied to agricultural products sold in rural markets, nor to those used at the household or small-scale level. Numerous reports, although still controversial, indicate that some mycotoxins may be degraded or inactivated during cereal fermentations. Steinkraus (1983), reported reduced aflatoxin B1 (AFB1) levels during Indonesian ontjom fermentations, while Van Veen et al. (1968) observed that the aflatoxin content of peanut press cake was reduced by both the ontjom mould, Neurospora and the tempe mould, R. oligosporus. Biological detoxification of AFB1 through degradation has, thus, far only been proven for Flavobacterium aurantiacum (Line et al., 1994; DSouza and Brackett, 1998), an organism which appears to be wrongly classified and which has now been identified as Nocardia corynebacterioides, a member of the Gram-positive phylogenetic group of the Actinomycetales (Holzapfel et al., unpublished data). Studies conducted under defined conditions using single LAB strains isolated from Ghanaian kenkey (aflata) resulted in reduced production of Alternaria toxin (Holzapfel, 1997). In addition, some authentic LAB strains and others isolated from fermented Turkish foods (L. plantarum and L. pentosus) were observed to reduce patulin concentrations by >60% in semisynthetic medium (Arici, 1997). 3.2.7. Biogenic amines Biogenic amines are frequently produced by amino acid decarboxylase positive microorganisms, during fermentation. The occurrence of biogenic amines in

traditional fermented foods has, however, been reported (Nout et al., 1994; Buckenhuskes et al., 1992). Certain LAB, such as L. buchneri, have been shown to produce biogenic amines, such as histamine, putrescine, tyramine and cadaverine, in fermented products of plant and animal origin. A number of bacteria associated with these fermentations, however, exhibit the potential for degrading histamine and tyramine through the production of mono- and di-amino-oxidases (Leuschner et al., 1998). 3.3. Experience with the application of starter cultures in small-scale operations Experience gained in the field of traditional fermentation technologies has shown that the addition of malted grains to fermentation media increases the rate of fermentation due to the endogenous amylolytic activity of the grains. Fermentation processes may also be accelerated through the addition of a starter obtained from a previous fermentation batch (backslopping). In traditional back-slopping, inoculum from a previous batch of fermented dough contains large numbers of desirable microorganisms in an active state, which are adapted to the substrate. Inocula consisting of a portion of a fermenting substrate may be preserved by dehydration (air- or sundrying) and grinding into a powder. Dehydration enhances the viability of microorganisms over relatively long periods, provided the product is maintained in the dehydrated state. Natural preservation of microorganisms can also be accomplished with the use of a carrier, such as the porous material of a gourd, fermen-

W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212

205

tation utensils or an inoculation belt, such as that used in Ghana for the initiation of pito beer fermentations. Although such preservation methodologies are common to many regions and probably have a long tradition, they have not been adequately studied. Product groups associated with traditional inoculation methodologies are summarised in Tables 5 8. Interesting examples of mixed-culture dough inocula prepared either in the form of dried powders, flat cakes or hard balls, are found in several Asian countries where they are used for the inoculation of starchy substrates in the production of alcoholic beverages. These starters contain mixed cultures of filamentous fungi (e.g. Amylomyces, Mucor, Rhizopus, Actinomucor), yeasts (Saccharomyces spp., Pichia spp., Hansenula spp.) and LAB (species of Lactobacillus and Pediococcus) (Tamang, 1998), and are referred to by different names in accordance with the location of production. These old traditions in starter preparation, preservation and distribution present an extremely valuables basis for the development and application of other types of starters in small-scale processing, and a

number of attractive challenges and benefits to the entrepreneur: (i) the use of simple, inexpensive utensils that are readily available; (ii) flexibility and simplicity of maintenance and handling; (iii) minimal losses due to fermentation failure; (iv) job creation and income generation in rural areas; (v) handling and storage at the household level; (vi) distribution and sale in local markets. 3.4. Small-scale vs. large-scale industrial applications; single- vs. mixed-strain starters Spontaneous fermentations typically result from the competitive activities of different microorganisms. Strains best adapted and with the highest growth rate dominant during particular stages of the process. The complexity and variability of microbial populations associated with these fermentations is somewhat reduced in back-slopping operations, where processing conditions and continued recycling of a portion of a previous batch, determines dominance of the best adapted strains. Inoculations with a single-strain culture can eventually result in a mixed-strain fermenta-

Table 5 Starter cultures used for traditional dairy products Product Dahi Country India and neighbour countries Raw materials buffalo milk, cow milk Amount of inoculum 1 2% (summer), 5 10% (winter) Storage/starter application lyophilised back-slopping Microorganisms Strains of one or more of: Lactococci, S. thermophilus, L. cremoris, L. delbruckii ssp. bulgaricus, L. acidophilus thermophilic LAB L. casei, L. plantarum, L. brevis, bacilli S. thermophilus and L. delbruckii ssp. bulgaricus or L. fermentum yeasts, L. brevis, L. kefiranofaciens

Tairu Kishk Kuschuk

Malaysia Egypt, Iraq, North Africa Greece, Cyprus, Turkey

cow milk, sojabeans cow milk, camel milk and wheat sheep milk and wheat

2.5 3.0% 1/3 (w/w)

back-slopping Laban zeer, dried kishk

Trahanas

Addition on a daily basis

back-slopping commercial cultures

Kefir

Russia (Caucasus)

goat milk, sheep milk, cow milk

25 30 g of kefir grains per 500 ml milk

kefir grains refrigeration

206

W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212

Table 6 Starter cultures applied in the preparation of traditional acid-leavened cereal and legume products Product Idli (Dosa) Country South India, Sri Lanka Raw materials rice, black gram, Catso wheat, maize, peas, cowpeas, soybeans rice Inoculum (starter) buttermilk, commercial yeast, dried idli Storage conditions refrigeration, dehydration Microorganisms L. mesenteroides, E. faecalis, S. cerevisiae, Torulopsis

Puto

Philippines

Injera (Ethiopia), Anjeira (Sudan)

Ethiopia, Sudan

tef (Ethiopia), sorghum (Sudan) or other cereals rice or wheat flour

Hopper (Appa)

Sri Lanka

(a) leba dura (ground slurry/ 18 h, sugar), (b) powdered puto irsho (ersho) and fermentation container (bohicka) bakers yeast or toddy (fermented drink) vigorously fermenting dough

(a) neutralisation of ground slurry, (b) dehydration cool place in home

L. mesenteroides, E. faecalis, S. cerevisiae

Candida and other yeast spp. LABa

Kisra

Sudan

sorghum

active fermentation

S. cerevisiae toddy: mixed culture years/LAB yeasts, LAB, acetic acid, bacteria

A majority of these products are manufactured by spontaneous fermentation. Examples given refer to alternative options. a Lactic acid bacteria.

tion if raw materials are not sterilised prior to inoculation and maintained axenic (free from foreign microorganisms) throughout strict process control. Single-strain cultures offer advantages of improving both process control and the predictability of me-

tabolic activities within the culture. They are, however, relatively easily degraded by bacteriophage infection, spontaneous mutation or through the loss of key physiological properties (e.g. plasmid-mediated fermentation of lactose). Deterioration in culture performance

Table 7 Mixed starter cultures applied in the production of traditional acid- and acid-alcoholic fermented gruels Product Mawe (sourdough) Mahewu Country Benin Raw materials Maize Inoculum (starter) previous batch small portion of wholewheat flour fermenting kocho Inoculation belt Storage conditions active fermentation (?) Microorganisms heterofermented lactobacilli and yeasts mainly heterofermented lactobacilli (yeasts) LABa yeasts Lactobacilli, yeasts

Southern Africa

Maize

Kocho (a flour) Pito (traditional beers)

Ethiopia Ghana, west Africa

ensete (false banana) Sorghum (maize millet)

pit (fermentation) drying

Most of these products are still manufactured by spontaneous fermentation. Examples given refer to alternative options. a Lactic acid bacteria.

W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212 Table 8 Ragi-type starter cultures Product Tape ketan Country Indonesia (JAVA) Himalayan regions (India, Nepal, Bhutan) Thailand China, Taiwan Nalyoia Malaysia Indonesia Raw materials rice Inoculum (starter) Ragi Storage conditions Dehydrated (air- or sundried)

207

Microorganisms different moulds, yeasts and LAB

Jaanr Chiang

rice millet, other cereals

Marcha, Bakhar, Ahab

Krachae Lao-Chao Tapai perlert Rice wine Tapai ubi tape kerccella

(bran) rice millet rice (millet) rice rice Baseam

Loogpang Chin-yueh ragi tapai ragi samsu Ragi

Sources: Steinkraus (1996), Tamang (1998) and Merican and Quee-Lan (1989).

due to one or more of these effects adversely affects the fermentation process. Modern equipment for the preparation, handling and application of pure singlestrain cultures and for strict process control at all stages of the fermentation, are not available or attainable in most small-scale operations. In large-scale fermentations, on the other hand, consistent end product quality is achieved over extended time periods through the use of defined single-strain cultures and properly controlled processes. Mixed strain cultures, such as those of the ragi type, on the other hand, are less susceptible to deterioration and are, thus, better suited to most smallscale operations. Mixed strain cultures are relatively unaffected by fluctuating conditions of handling, storage and applications. In addition, they contribute to a more complex sensory quality, whilst producing favourable synergistic effects, such as the degradation of undesirable factors, flavour production and accelerated ripening and maturation. Variation in product quality with the use of mixed strain cultures can be minimised through proper process control. Although moulds play a minor role in the fermentation of foods in Africa, they are of major importance in Asian food fermentations. In Europe, traditional mould-ripened foods are mainly restricted to bluemould (Penicillium roqueforti) and white-mould (P. camemberti) cheeses and mould-ripened fermented

sausages (containing either P. nalgiovense or P. chrysogenum) (Geisen, 1993). Plant materials containing fermentable sugars provide suitable substrates for yeast species of Saccharomyces, Candida, Torula, Hansenula and others. These yeasts, in particular Saccharomyces spp., are typically associated with spontaneous alcoholic fermentations, such as in the production of African opaque beers, palm wine and Asian types of beverages, such as rice wines, palm wines and Indonesian tape. Spontaneous alcoholic fermentation of palm wines and most opaque cereal beers. Selected strains of S. cerevisiae are used for the industrial production of both western style and traditional African beers. Dehydrated yeast, which is mainly applied in bread making, is readily available on the market throughout Africa. It is also applied in small-scale beer brewing (Tables 6 and 7). LAB are major importance among bacteria associated with traditional fermented foods. The largest spectrum and richest variety of lactic fermented foods is probably found in Africa. The association of LAB with the human environment and their beneficial interactions, both in food and in the human intestinal tract, combined with the long tradition of lactic fermented foods in many cultures, have led to the conclusion that these foods may be generally recognised as safe (GRAS). Several factors must be taken into consideration when evaluating the use of LAB as starters

208

W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212

(Holzapfel, 1997). (i) Not all LAB are of equal technical and practical importance in food fermentations. (ii) Lactobacillus (both homo- and heterofermentative), Leuconostoc and, to a lesser extent, Pediococcus, Lactococcus, Enterococcus and Weissella are the genera which generally occur in traditional fermented foods. (iii) The genus Bifidobacterium, although phylogenetically not related to LAB, is often grouped as part of the LAB for its probiotic functions (Reddy and Rivenson, 1993; Salminen et al., 1996; Holzapfel et al., 1998). (iv) With the exception of S. thermophilus, species of the genus Streptococcus are generally regarded as pathogens. (v) Suitable cultures for fermentation must be selected at the strain level since not all strain of a species are equally suitable for use as starters, nor are all equally well adapted to a food substrate. (vi) A number of industrial lactic food fermentations (e.g. the production of sauerkraut and dill cucumbers) are spontaneous although technically well controlled. The association of certain strains of Enterococcus faecium, E. faecalis and L. rhamnosus, with exceptional cases of endocarditis, should be no absolute reason to disqualify the use of food grade strains of these and other LAB species from their potential use in food fermentations and even as probiotics. LAB starter cultures are not yet commercially available for the small-scale fermentation of traditional African foods. Making them available to the small-scale processor, however, poses a great challenge to both the food microbiologist and the potential entrepreneur.

4. Handling, maintenance and distribution of starter cultures for small-scale fermentations 4.1. Traditions in the conventional handling, preservation and application of starter cultures Perhaps the oldest traditions in the preparation, handling and distribution of starter cultures are to be found in the different regions of Asia (Lee and Fujio, 1999). This is particularly true for the mixed-culture dough inocula, such as the ragi-type starter cultures which have been used for centuries in the production of a variety of sweet and sour alcoholic beverages and pastes (Steinkraus, 1997; Tamang, 1998) (Table 8). Although ragi production does not incorporate the use

of specialised equipment, ragi formulations are maintained proprietary by manufacturers. Powdered ragi from a previous batch is sprinkled as an inoculum over the paste prepared from rice flour and water and moulded into a ball. Inoculated balls are placed on bamboo trays and either covered with muslin cloth (Malaysia) or with ferns (Himalayas). Microbial growth takes place over a 2 5-day period under ambient conditions, during which gradual desiccation of the rice balls occurs. Slow drying of ragi balls during the rainy season, results in large numbers of Mucor and Rhizopus spp. (Merican and Quee-Lan, 1989). In some countries of the Near East (Egypt, Iraq) and North Africa dried kishk or laban beer is used as an inoculum for kishk and kuschuk production (Table 5). Relatively little information is available on starter culture traditions in Sub-Saharan Africa. The use of back-slopping approaches for inoculation are, however, widespread in that region. One example of a preserved starter is the inoculation belt, typical of Ghana and some countries in West Africa. The inert surface of the belt or woven rope, which consists of flax of hennep, facilitates the preservation of essential microorganisms during drying and storage. Sundrying may destroy some microorganisms and thereby reduce viable numbers, while slow and insufficient airdrying during the rainy season may result in contamination and poor quality starters. The shelf life of dehydrated starters may be enhanced by storage in an airtight container. Bakers yeast is used worldwide in bread baking. It is also applied in brewing and the production of wine at the household level. Bakers yeast is commonly used in the fermentation of sorghum and other cereal beers in Africa (Holzapfel, 1989). The predominance of lactic food fermentations in Africa and the contribution of these fermentations to food safety deserve special attention. Low expected turnover, distribution over extensive areas, and sophisticated methods for propagation and preservation are factors which complicate LAB starter culture development. Studies into the application of selected LAB strains for traditional fermentations, however, suggest sufficient potential for the use of LAB starters in small-scale food fermentations in Africa (Holzapfel, 1997). (i) The traditional starter kudeme is used as an inoculant in West Africa, for preparing the fermenting cassava dough, agbelima (Amoa-Awua and Ja-

W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212

209

kobsen, 1996). (ii) A stable LAB-enriched starter dough, aflata applied in the preparation of the fermented maize-meal product, kenkey, produces accelerated acidification and contains a natural selection of acid resistant strains (Nout et al., 1989; Nche et al., 1994). Cabinet and drum-drying of this dough from a 54% to a 10% moisture content, was shown to enhance the viability of LAB strains (Nche et al., 1994), thereby indicating the potential for its extended storage and distribution as a starter. (iii) Addition of CaCO3 (chalk) to fermenting substrates increases retention of the metabolic activities of LAB. CaCO3 confers this protective effect primarily through the neutralisation of lactic acid. It is not, however, effective in inhibiting undesirable microbes, such as yeasts and moulds. (iv) fermenting L. plantarum, isolated from kivunde, a traditional product of the Western and North-Western Regions of Tanzania produced safe cyanide levels < 10 mg/kg during cassava fermentations (Kimaryo et al., 2000). Kivunde is typically formed into small balls, (3 5 cm in diameter) and air-dried, thereby preserving viable strains over extended periods. These starter cultures may find application and might serve to improve small-scale fermentations even in rural areas, upon identification and selection of suitable strains. Such approaches will, however, only be successful if the basic principles of good processing practices (GMP) are observed. 4.2. Logistical requirements, pragmatic approaches and economic considerations Apart from experience with the traditional use of mixed culture inoculants of the ragi type in Asia, relatively little is known about the economics and practical constraints of starter culture supply and distribution to small-scale processors in developing countries. Infrastructure required for the manufacture, distribution and storage (e.g. by refrigeration) of starters on a continuous basis is generally available in urban areas. This is not however the case in most rural areas of developing countries. Information on the logistics of starter culture distribution in developing countries is, therefore, required. Research priorities for starter culture development were extensively reviewed elsewhere in this publication (Motarjemi and Asante, 2002). Four major focal areas for research

on starter cultures were recommended by WHO and FAO (FAO/WHO, 1996). (i) Assessment of the need for, and feasibility of, using starter cultures. (ii) Establishment of an appropriate level of starter culture technology. (iii) Development of appropriate starter culture delivery mechanisms. (iv) Selection of microorganisms with desirable properties.

5. Approaches toward starter cultures with improved properties 5.1. Multifunctional considerations On the basis of all of the foregoing discussion, it is clear that the selection of starter cultures should take account of properties beyond those of acid (by LAB) or alcohol production (by yeasts). Prospects for improving the safety, shelf life, sensory characteristics, nutritional value and even health-promoting properties of fermented foods by multifunctional starter cultures, have become a reality in our time. Modern selection techniques (Zhong et al., 1998) and the application of molecular biological tools have been covered elsewhere in this publication (Valyasevi and Rolle, 2002). These tools and techniques provide improved means for obtaining the most suitable microbial strains for each specific substrate and situation and have opened up opportunities for the development of tailor-made starter cultures. 5.2. Genetic improvements: pro and contra It is doubtful whether the most modern techniques and selection procedures would result in a multifunctional strain having all desirable metabolic features. Selected strains may be improved through the application of genetic technologies. Recombinant DNA technology may be applied in the production of tailor-made starter cultures which would meet technical and metabolic requirements necessary for a specific fermentation (e.g. accelerated acid production, improved wholesomeness, health-promoting properties, overproduction of bacteriocins or of particular enzymes necessary for degradation of undesirable factors). Undesirable properties, such as mycotoxin or antibiotic production by food-grade moulds, may be eliminated by techniques, such as gene disruption (Geisen and

210

W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212

Holzapfel, 1996; Hammes and Vogel, 1990). A large number of such optimised cultures already exist. Regulatory issues however preclude their use.

6. Conclusions Knowledge on traditional fermentations is rapidly increasing. Studies on microbial dynamics, substraterelated interactions and metabolic activities of different microbial groups, on key enzymes and on the role of technical and other process parameters, have provided a firm basis for improvement of traditional, small-scale and household level fermentation processes. In addition, an understanding and application of HACCP principles and the observance of GMP are of vital importance. The introduction of appropriate starter culture techniques may constitute one major step towards improved safety, quality and security of traditional small-scale fermentation. Artisanal level starter culture traditions in Asia have proven feasible over generations, and may serve as model for application in other regions. Focused studies toward the introduction of starter cultures for small-scale fermentations seem more than justified, and in fact deserve the highest priority. A number of aspects of relevance to starter culture production should, therefore, receive special attention. As examples, the following should be mentioned. (i) Information on inhibition kinetics of food-borne pathogens by typical LAB cultures under practical, product-specific conditions. (ii) Viability and survival of sublethally injured enteropathogens in fermented foods and their potential pathogenicity. (iii) The role of bacteriocin producing LAB strains as an additional safety factor against Gram-positive pathogens (e.g. S. aureus) and improvement of inhibitory action, e.g. by synergistic effects among different antimicrobial agents and physico-chemical factors related to the food substrate. (iv) information on the typical concentrations of D( )-lactic acid in traditional fermented gruels and beverages (e.g. traditional beer types), the average daily intake of such beverages, and the relevance of D( )-lactic acid producing LAB strains in foods typically consumed in large quantities. (v) Growth patterns, stability and acidification potential of L(+)-lactate producing strains during smallscale operations. (vi) Amino acid decarboxylase activ-

ity as a negative selection feature for potential starter cultures, and mono-amino-oxidase activities as a positive selection criterion, with the aim of reducing levels of biogenic amines in selected fermented foods. (vii) Potential for the fermentative detoxification of mycotoxins, approaches for strain selection and toxicological studies on possible degradation products. (viii) Studies on functional (probiotic) properties of LAB strains involved in traditional fermented foods, development of in vitro selection methods for probiotic strains, and studies on growth behaviour and technical performance of probiotic strains.

Acknowledgements Some data presented have been generated from research projects that were financially supported by the German Ministry of Nutrition, Agriculture and Forestries (BMVEL) and the European Commission (STD projects TS2.0267 and TS3 * CT-940344). These are gratefully acknowledged.

References
Abudu, I.A., Akinyele, I.O., 1990. The effect of germination on the oligosaccharides, trypsin inhibitors and nutrient content of cowpea milk. Food Chem. 35, 161 166. Adegoke, G.O., Otumu, E.J., Akanni, A.O., 1994. Influence of grain quality, heat and processing time on the reduction of aflatoxin B, levels in tuwo and ogi: two cereal-based products. Plant Foods Hum. Nutr. 45, 113 117. Aderibigbe, E.Y., Odunfa, S.A., 1990. Growth and extracellular enzyme production by strains of Bacillus species isolated from fermenting African locust bean, Iru. J. Appl. Bacteriol. 69, 662 671. Ahrne, S., Molin, G., 1991. Spontaneous mutations changing the raffinose metabolism of Lactobacillus plantarum. Antonie Van Leeuwenhoek 60, 87 93. Amoa-Awua, W.K., Jakobsen, M., 1995. The role of Bacillus species in the fermentation of cassava. J. Appl. Bacteriol. 79, 250 256. Amoa-Awua, W.K., Jakobsen, M., 1996. The role of microorganisms in the fermentation of agbelima cassava dough. In: Halm, M., Jakobsen, M. (Eds.), Traditional Fermented Food Processing. The Third Biennial Seminar on African Fermented Food, Ghana, July 1996. WAITRO, Taastrup, Denmark, pp. 7 15, ISBN 87-87047-23-3. Amoa-Awua, W.K., Appoh, F.E., Jakobsen, M., 1996. Lactic acid fermentation of cassava dough into agbelima. Int. J. Food Microbiol. 31, 87 98.

W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212 Amoa-Awua, W.K., Halm, M., Jakobsen, M., 1998. HACCP System For Traditional African Fermented Foods: Kenkey. WAITRO, 1998, Taastrup, Denmark. ISBN: 87-90737-02-4. Arici, M., 1997. Untersuchungen an Milchsaurebakterien zum fer mentativen Abbau von Patulin. Doctorate thesis, Technische Universitat Karlsruhe, Germany. Buckenhuskes, H.J., 1993. Selection criteria for lactic acid bacteria to be used as starter cultures for various food commodities. FEMS Microbiol. Rev. 12, 253 272. Buckenhuskes, H.J., Sabatke, I., Gierschner, K., 1992. Zur frage des vorkommens biogener amine in milchsauer fermentiertem gemuse. Critical review. Industrielle Obstrund Gemuseverwertung 7/92, 255 263. Chavan, J.K., Kadam, S.S., 1989. Nutritional improvement of cereals by fermentation. Crit. Rev. Food Sci. Nutr. 28, 349 400. DSouza, D.H., Brackett, R.E., 1998. The role of trace metal ions in aflatoxin B1 degradation by Flavobacterium aurantiacum. J. Food Prot. 32, 1666 1669. Ejiofor, M.A.N., Okafor, N., 1981. Comparison of pressed and unpressed cassava pulp for gari making. In: Terry, E.R., Oduro, K.A., Caveness, F. (Eds.), Tropical Root Crops: Research Strategies for the 1980s. IDRC, Ottawa, Canada, pp. 154 158. Essers, A.J.A., Jurgens, C.M.G.A., Nout, M.J.R., 1995. Contribution of selected fungi to the reduction of cyanogen levels during solid substrate fermentation of cassava. Int. J. Food Microbiol. 26, 251 257. FAO/WHO, 1996. Fermentation: assessment and research. Report of a Joint FAO/WHO Workshop on Fermentation as a Household Technology to Improve Food Safety. Pretoria, South Africa, 11 15 December 1995. WHO/FNU/FOS/96.1. Geisen, R., 1993. Fungal starter cultures for fermented foods: molecular aspects. Trends Food Sci. Technol. 4, 251 256. Geisen, R., Holzapfel, W.H., 1996. Genetically modified starter and protective cultures. Int. J. Food Microbiol. 30, 315 324. Graham, G.G., MacLean Jr., W.C., Morales, E., Hamaker, B.R., Kireis, A.W., Mertz, E.T., Axtell, J.D., 1986. Digestibility and utilisation of protein and energy from nasha, a traditional Sudanese fermented sorghum weaning food. J. Nutr. 116, 978 984. Hahn, S.K., 1989. An overview of African traditional cassava processing and utilization. Outlook Agric. 18, 110 118. Hamaker, B.R., Kirleis, A.W., Butler, L.G., Axtell, J.D., Mertz, E.T., 1987. Improving the in vitro protein digestibility of sorghum with reducing agents. Proc. Natl. Acad. Sci. 84, 626 628. Hammes, W.P., 1990. Bacterial starter cultures in food production. Food Biotechnol. 4, 383 397. Hammes, W.P., Hertel, C., 1998. New developments in meat starter cultures. Meat Sci. 49 (Suppl. 1), S125 S138. Hammes, W.P., Vogel, R.F., 1990. Gentechnik zur Modifizierung von Starterorganismen. Lebensmitteltechnik 1 2, 24 32. Holzapfel, W.H., 1989. Industrialization of mageu fermentation in South Africa. In: Steinkraus, K.H. (Ed.), Industrialisation of Indigenous Fermented Foods. Marcel Dekker, New York, pp. 285 328. Holzapfel, W.H., 1997. Use of starter cultures in fermentation on a household scale. Food Control 8, 241 258. Holzapfel, W.H., Geisen, R., Schillinger, U., 1995. Biological preservation of foods with reference to protective cultures, bacter-

211

iocins and food-grade enzymes. Int. J. Food Microbiol. 24, 343 362. Holzapfel, W.H., Haberer, P., Snel, J., Schillinger, U., Huis int Veld, J.H.J., 1998. Overview of gut flora and probiotics. Int. J. Food Microbiol. 41, 85 101. Hounhouigan, D.J., Nout, M.J.R., Nago, C.M., Houben, J.H., Rombouts, F.M., 1994. Starter cultures of lactobacilli and yeasts in the fermentation of mawe porridge. In: Hounhouigan, D.J. (Ed.), Fermentation of Maize (Zea mays L.) Meal for mawe Production in Benin. Doctorate thesis, Agricultural University of Wageningen, Wageningen, the Netherlands. Hounhouigan, D.J., Nout, M.J.R., Nago, C.M., Houben, J.H., Rombouts, F.M., 1999. Use of starter cultures of lactobacilli and yeast in the fermentation of mawe, an African maize product. Trop. Sci. 39, 220 226. Kakade, M.L., Rackis, J.J., McGhee, J.E., 1974. Determination of trypsin inhibitor activity of soy products. Cereal Chem. 51, 376 382. Kazanas, N., Fields, M.L., 1981. Nutritional improvement of sorghum by fermentation. J. Food Sci. 44, 820. Khetarpaul, N., Chauhan, B.M., 1989. Effect of fermentation by pure cultures of yeasts and lactobacilli on phytic acid and polyphenol content of pearl millet. J. Food Sci. 54, 780 781. Kimaryo, V.M., Massawe, G.A., Olasupo, N.A., Holzapfel, W.H., 2000. The use of starter cultures in the fermentation of cassava for the production of kivunde, a traditional Tanzanian food product. Int. J. Food Microbiol. 56, 179 190. Lee, C.-H., 1997. Lactic acid fermented foods and their benefits in Asia. Food Control 8, 259 269. Lee, A.C., Fujio, Y., 1999. Microflora of banh men, a fermentation starter from Vietnam. World J. Microbiol. Biotechnol. 15, 57 62. Leuschner, R.G., Heidel, M., Hammes, W.P., 1998. Histamine and tyramine degradation by food fermenting microorganisms. Int. J. Food Microbiol. 39, 1 10. Line, E.J., Brackett, R.E., Wilkinson, R.E., 1994. Evidence for degradation of aflatoxin B1 by Flavobacterium aurantiacum. J. Food Prot. 57, 788 791. Lopez, Y., Gordon, D.T., Fields, M.L., 1983. Release of phosphorus from phytate by natural lactic fermentation. J. Food Sci. 48, 953 954. Lorri, W.S.M., 1933. Nutritional and Microbiological evaluation of Fermented Cereal Weaning Foods. PhD thesis, Chalmers University of Technology, Goteborg, Sweden. Mahajan, S., Chauhan, B.M., 1987. Phytic acid and extractable phosphorus of pearl millet flour as affected by natural lactic acid fermentation. J. Sci. Food Agric. 41, 381 386. Mbugua, S.K., 1986. The nutritional and fermentation characteristics of uji from dry milled maize flour (Unga Baridi) and whole wet milled maize. Food Chem., Microbiol., Technol. 10, 154 161. Mbugua, S.K., Ahrens, R.H., Kigutha, H.N., Subramanian, V., 1992. Effect of fermentation, malted flour treatment and drum drying on nutritional quality of uji. Ecol. Food Nutr. 28, 271 277. Merican, Z., Quee-Lan, Y., 1989. Tapai processing in Malaysia: a technology in transition. In: Steinkraus, K.H. (Ed.), Industrialisation of Indigenous Fermented Foods. Marcel Dekker, New York, pp. 169 188. Milliere, J.B., Mathot, A.G., Schmitt, P., Divies, C., 1989. Pheno-

Edited by Foxit Reader Copyright(C) by Foxit Software Company,2005-2007 For Evaluation Only.
212 W.H. Holzapfel / International Journal of Food Microbiology 75 (2002) 197212 Report of the 8th Bean Research Conference, Mellaire, MI, ARS 74-41, 68 71. Smith, J.E., Lewis, C.W., Anderson, J.G., Solomons, G.L., 1994. Mycotoxins in human nutrition and health. European Union Directorate-General XII, Report 16048EN. Sorensen, W.G., Hesseltine, C.W., 1966. Carbon and nitrogen utilization by Rhizopus oligosporus. Mycologia 58, 681 689. Steinkraus, K.H., 1983. Handbook of Indigenous Fermented Foods. Marcel Dekker, New York. Steinkraus, K.H., 1996. Handbook of Indigenous Fermented Foods, 2nd edn. Marcel Dekker, New York Revised and Expanded. Steinkraus, K.H., 1997. Classification of fermented foods: worldwide review of household fermentation techniques. Food Control 8, 311 317. Svanberg, U., Sandberg, A.-S., 1988. Improving iron availability in weaning foods through the use of germination and fermentation. In: Aldwick, D., Moses, S., Schmidt, O.G. (Eds.), Improving young child feeding in Eastern and Southern Africa: Household Level Food Technology. Proc., Workshop, Nairobi, Kenya, October 1987. IDRC-265e, Ottawa, Canada, pp. 366 373. Svanberg, U., Lorri, W., Sandberg, A.-S., 1993. Lactic fermentation of non-tannin and high-tannin cereals: effects on in vitro estimation of iron availability and phytate hydrolysis. J. Food Sci. 58, 408 412. Tamang, J.P., 1998. Role of microorganisms in traditional fermented foods. Indian Food Ind. 17, 162 167. Trugo, L.C., Ramos, L.A., Trugo, N.M.F., Souza, M.C.P., 1990. Oligossacharide composition and trypsin inhibitor activity of P. vulgaris and the effect of germination on the alpha-galactoside composition and fermentation in human colon. Food Chem. 36, 53 61. Valyasevi, R., Rolle, R.S., 2002. An overview of small-scale food fermentation technologies in developing countries, with special reference to Thailand: scope for their improvement. Int. J. Food Microbiol. 75, 231 239. Van Veen, A.G., Graham, D.C.W., Steinkraus, K.H., 1968. Fermented peanut presscake. Cereal Sci. Today 13, 96 99. Westby, A., Choo, B.K., 1994. Cyanogen reduction during the lactic fermentation of cassava. Acta Hortic. 376, 209 215. Westby, A., Reilly, A., Bainbridge, Z., 1997. Review of the effect of fermentation on naturally occurring toxins. Food Control 8, 329 339. WHO, 1968. Food Additives 29, 144 148. WHO, 1995. Hazard analysis critical control point system: concept and application. Report of a WHO Consultation with the Participation of FAO, 29 31 May 1995. WHO/FNU/FOS/95.7. Zhong, W., Millsap, K., Bialowska-Hobrzanska, H., Reid, G., 1998. Differentiation of Lactobacillus species by molecular typing. Appl. Environ. Microbiol. 64, 2418 2423.

typic characterisation of Leuconostoc species. J. Appl. Bacteriol. 67, 529 542. Mital, B.K., Schallenberger, R.S., Steinkraus, K.H., 1973. Alphagalactosidase activity of lactobacilli. Appl. Microbiol. 26, 783 788. Motarjemi, Y., Asante, A., 2002. Impact of small scale fermentation technology on food safety in developing countries. Int. J. Food Microbiol. 75, 213 229. Nche, P.F., 1995. Innovation in the production of kenkey, a traditional fermented maize product of Ghana. Nutritional, physical and safety aspects. PhD thesis, Agricultural University of Wageningen, the Netherlands. Nche, P.F., Nout, M.J.R., Rombouts, F.M., 1994. The effect of cowpea-supplementation on the quality of kenkey, a traditional Ghanaian fermented maize food. J. Cereal Sci. 19, 191 197. Nout, M.J.R., Rombouts, F.M., Havelaar, A., 1989. Effect of accelerated natural lactic fermentation of infant food ingredients on some pathogenic microorganisms. Int. J. Food Microbiol. 8, 351 361. Nout, M.J.R., Nche, P.F., Hollman, P.C.H., 1994. Investigation of the presence of biogenic amines and ethyl carbamate in kenkey made from maize and maize cowpea mixtures as influenced by process conditions. Food Addit. Contam. 11, 397 402. Ogun, P.O., Markakis, P., Chenoweth, W., 1989. Effect of processing on certain antinutrients in cowpeas (Vigna unguiculata). J. Food Sci. 54, 1084 1085. Olasupo, N.A., Olukoya, D.K., Odunfa, S.A., 1994. Plasmid profiles of bacteriocin-producing Lactobacillus isolates from African fermented foods. Folia Microbiol. 39, 181 186. Olasupo, N.A., Olukoya, D.K., Odunfa, S.A., 1997. Identification of Lactobacillus species associated with selected African fermented food. Zeitschrift fur Naturforschung C.-J. Biosci. 52, 105 108. Oyewole, O.B., 1997. Lactic fermented foods in Africa and their benefits. Food Control 8, 289 297. Reddy, B.S., Rivenson, A., 1993. Inhibitory effect of Bifidobacterium longum on colon, mammary and livery carcinogenesis induced by 2-amino-3-methylimidazo[4,5-f]quinoline, a food mutagen. Cancer Res. 53, 3914 3918. Reddy, M.R., Pierson, M.D., Salunkhe, D.K., 1986. Legume-Based Fermented Foods. CRC Press, Boca Raton, FL. Salminen, S., Isolauri, E., Salminen, E., 1996. Clinical uses of probiotics for stabilizing the gut mucosal barrier: successful strains for future challenges. Antonie van Leeuwenhoek 70, 347 358. Schillinger, E., Geisen, R., Holzapfel, W.H., 1995. Bakteriozine von milchsaurebakterien-eigenschaften, wirkungsmechanismen und genetik. Biospektrum 1, 59 66. Shallenberger, R.S., Hand, D.B., Steinkraus, K.H., 1967. Changes in sucrose, raffinose and stachyose during tempeh fermentation.