~w~l ELSEVIER Journal International Journal of Food Microbiology 34 (1997) 249-258 Munkoyo beverage, a traditional Zambian fermented maize gruel using Rhynchosia root as amylase source R.M. Zulu 1, V.M. Dillon, J.D. Owens* Department of Food Science and Technology, The University of Reading, RG6 6AP Reading, UK Received 13 March 1996; revised 17 August 1996; accepted 3 October 1996 Abstract A typical munkoyo beverage was made by fermenting Rhynchosia heterophylla root extract-cooked maize meal mixture with Lactobacillus confusus LZ1 and Saecharomyces cerevisiae YZ20. The fermented munkoyo beverage had a pH of 3.3, lactic acid content of 60 mmol/l, ethanol 320-410 mmol/1 and a characteristic 'munkoyo' aroma. L. confusus, used alone, produced a beverage with a faint munkoyo flavour note whilst beverage produced with S. cerevisiae alone seemed not to have a typical munkoyo note. R. heterophylla root extract converted 75% of the starch in sterile cooked maize meal to maltose (80% of total sugars), maltotriose (17%) and glucose (3%) in 1 h at 45°C. During fermentation by the mixed culture or the yeast alone most of the maltose was utilised but little or none of the maltotriose. The ratio of yeast to lactic acid bacteria in the starter culture affected the rate of production of ethanol but had no effect on the growth or acid production by the bacterium. To obtain a munkoyo beverage with the desired low alcohol concentration ( < 100 mmol/1), the ratio of yeast concentration (cfu/ml) to Lactobacillus concentration in the starter culture should be 1:1000 or less and the beverage should be fermented for 24 h only. Copyright © 1997 Elsevier Science B.V. Keywords: Amylase; Lactobacillus; Maize; Munkoyo beverage; Rhynchosia; Saccharomyces * Corresponding author. Tel.: +44 1189 316758; fax: +44 1189 310080. Present address: National Council for ScientificResearch, Food Technology Research Unit, P.O Box 310 158, Lusaka, Zambia. 0168-1605/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved PH S0168-1605(96)01 195-6 250 R.M. Zulu et al. / Int. J. Food Microbiology 34 (1997) 249-258 1. Introduction Many yeasts and most lactic acid bacteria lack amylolytic enzymes and, hence, fermentation of starchy substrates requires the addition of an amylase source, such as malt or koji. Munkoyo beverage is a traditional cereal-based fermented beverage, prepared in Zambia and the Shaba area in Zaire (Simwamba and Elahi, 1986; Pauwels et al., 1992; Steinkraus, 1996). Roots ofEminia, Rhynchosia and Vigna species, known locally in Zambia under the generic name of munkoyo, are used as the source of amylase in this beverage (Mulkay et al., 1985). The roots contain exceptionally high levels of ~- and fl-amylases (Mulkay et al., 1985) and effect a rapid liquefaction of the porridge gel. The fermentation is due to naturally occurring microorganisms, primarily lactic acid bacteria and yeasts, and occurs at ambient temperatures (25-30°C) in 24-48 h. The fermented beverage is sweet-sour (pH 3.5), with less than 5 g ethanol/kg, but becomes more alcoholic (14-26 g/kg) if it is fermented for a longer period of time (Steinkraus, 1996). The beverage has a characteristic flavour which is referred to as 'munkoyo' flavour and is consumed daily by all family members, but particularly by women and children (Steinkraus, 1996). Munkoyo beverage is mainly prepared in households and is consequently of varying quality. Munkoyo beverage has similarities to mahewu, a traditional South African fermented maize beverage, except that in the preparation of mahewu wheat flour and/or sorghum malt is used as a source of amylolytic enzymes (Holzapfel, 1989). Traditionally fermented mahewu is dominated by heterofermentative lactic acid bacteria but mahewu manufactured industrially generally uses pure cultures of thermophilic homofermentative lactic acid bacteria. This yields a clean lactic acid flavour, less piquant than that produced by fermentations involving heterofermentative lactic acid bacteria and yeasts (Holzapfel, 1989). The major aim of this study was to examine the roles of the lactic acid bacteria and yeasts in the fermentation of munkoyo beverage, particularly in relationship to the development of typical munkoyo aroma. A second aim was to identify the substrates used, the metabolites produced and to determine the nature of any interactions between the microorganisms. 2. Materials and methods 2.1. Organisms Lactobacillus confusus LZ1, a heterofermentative species, and Saccharomyces cerevisiae YZ20 were isolated from fermented Rhynchosia root extract-maize meal mixture after serial enrichment through 10 successive subcultures (Zulu, 1993). In mixed culture they yield a fermented product having the characteristic munkoyo aroma. L. confusus was maintained in APT-chalk-semi-solid agar medium and stored at 5°C (Nuraida et al., 1995). S. cerevisiae was maintained on Potato dextrose agar slopes (Oxoid CM139; Unipath, Basingstoke RG24 8PW, UK). R.M. Zulu et al. / Int. J. Food Microbiology 34 (1997) 249 258 251 2.2. Maize meal and Rhynchosia roots White roller-grade maize meal was obtained from National Milling Company, Lusaka, Zambia. It had a moisture content of 117 g/kg and a starch content of 820 g/kg dry weight (determined according to Graffham, 1994). Roots of R. heterophylla were collected in the Copperbelt province of Zambia. The roots were debarked immediately after collection, beaten into threads, sun dried and then dispatched to Great Britain. The roots and maize meal were stored at room temperature (~, 18°C) until required. 2.3. Preparation o f sterile root extract The roots were ground in a hammer mill (203 mm diameter, screen 137 with 3.175 mm diameter holes; Christy and Norris, Chelmsford CM1 1SA, UK). Ground root, 200 g, was added to 1 1 warm (45-50°C) water and soaked for 1 h. The extract was then filtered successively through Whatman No. 4 filter paper (Whatman, Maidstone ME16 0LS, UK), nylon mesh (pores 5 x 5 pm; Wilson Sieves, Nottingham, UK) and a membrane filter (0.45 /tm diameter pores; Whatman 6780-2504). 2.4. Preparation o f sterile Rhynchosia root extract-maize meal mixture Maize meal (380 g) was added to 3800 ml warm water (40-45°C), heated to boiling with constant stirring and simmered for about 1 h with occasional stirring. The porridge was sterilised by autoclaving at 121°C for 5 min and then cooled to 45°C, during which time it formed a semi-solid mass. Sterile Rhynchosia root extract was added (280 ml-2.8 1 porridge) and the mixture was thoroughly stirred with a sterile glass rod. The mixture was held at 45°C until it was completely liquified ( ~ 1 h). 2.5. Determ&ation o f concentrations o f viable microbes Samples were diluted as required in 0.1% peptone water (Oxoid CM9). Counts of aerobic mesophilic bacteria were made on duplicate spread plates of Plate count agar (Oxoid CM325) incubated at 30°C for 24 h. Colonies 2 mm or more in diameter were counted. These are presumed to represent the aerobic microflora and to exclude lactic acid bacteria. Mesophilic aerobic bacteria were less than 10 cfu/ml in all the beverages at all times. Lactic acid bacteria were enumerated by the pour-plate method on duplicate plates of MRS agar (Oxoid CM361). Plates were overlayed with sterile medium and incubated at 30°C for 3 days. Yeasts were enumerated on duplicate, surface inoculated plates of acidified Potato dextrose agar (Oxoid CM139; pH adjusted to 3.5 by adding 1 ml of 10% (w/v) lactic acid solution to 100 ml molten medium) at 30°C for 3 days. 252 R.M. Zulu et al. / Int. J. Food Microbiology 34 (1997) 249-258 In general, all samples were examined for aerobic mesophiles, lactic acid bacteria and yeasts. For cultures inoculated with a single organism this served to demonstrate the absence of contaminating organisms, and for sterile control samples it demonstrated the absence of any microbes. 2.6. Analysis o f sugars, acids and ethanol by H P L C Glucose, maltose, maltotriose, acetic acid, lactic acid and ethanol were determined by HPLC on an Aminex HPX 87H column (Bio-Rad, Richmond, CA 94804, USA) at 45°C with a cation H + guard column (Bio-Rad 125-0129) and eluted with degassed (by bubbling helium) 0.005 mol/1 HzSO 4 in high purity water (Purite RO50 with Still Plus HP unit, conductivity < 0.08 /iS/cm; Purite, High Wycombe, UK) at a flow rate of 0.6 ml/min. The HPLC apparatus (Thermo Separation Products, Fremont, CA 94537, USA) comprised an Isochrom isocratic pump, a model 100 UV detector at 210 nm, an 8430 refractive index detector, an SP4400 dual channel integrator and an Altex 210 injector (Beckman Instruments, San Ramon, CA 94583, USA) with a 50 /tl loop. Concentrations were determined from standard curves relating peak area to concentration. Frozen samples were thawed, filtered through Whatman No. 4 filter paper and Millipore Ultrafilters (5000 NMWL, Cat No. UFC-LCC25; Millipore UK, Watford WD1 8YW, UK). Culture filtrates were diluted fifty times in 0.0025 mol/1 HzSO 4 immediately before injecting into the column. Samples of sterile mixtures were diluted only as necessary. Duplicate injections were done for all samples. Undiluted filtered samples were stored, if necessary, at - 1 8 ° C before HPLC analysis. A solution of sorbitol (1.0 mmot/1) was used as a standard to check the performance of the HPLC and was chromatographed after every six runs. The results were acceptable when the calculated concentration of sorbitol was 1.0 _+ 0.05 mmol/1. The identity of the maltose peak was confirmed by demonstrating that treatment with ~-glucosidase (Sigma M3145; Sigma-Aldrich, Poole, BH12 4QH, UK) completely converted the maltose peak to glucose and that treatment with invertase (Sigma 1-4504) had no effect. 2. 7. Amylolysis o f maize porridge by Rhynchosia root extract Sterile root extract (90 ml) was added to each of three beakers containing 900 ml maize meal porridge at 45°C. The beakers were incubated at 45°C for 1 h and then at 30°C for 48 h. Samples (10 ml) were removed from each after 1, 3, 5, 7, 10, 15, 20, 30 and 60 min and 7, 13, 25 and 49 h. Sodium hydroxide solution (1 ml of 0.5 mol/1) was immediately added to stop enzyme activity (Mulkay et al., 1985) and samples were assayed for sugars by HPLC. R.M. Zulu et al. / Int. J. Food Microbiology 34 (1997) 249 258 253 2.8. Preparation of starter cultures L. confusus LZ1 was streaked onto MRS agar and incubated anaerobically (Oxoid gas generating kit, BR 38) at 30°C for 3 days. The growth was harvested with a sterile dry cotton swab which was then immersed in 4 ml of sterile purified water (Purite RO50 reverse osmosis and HP deionisation unit). The suspension was added to 100 ml of hydrolysed porridge and incubated at 30°C for 24 h. S. cerevisiae YZ20 was streaked onto Potato dextrose agar and incubated at 300C for 3 days. The culture was harvested with a loop and a dense suspension was prepared in 4 ml of sterile purified water. The suspension was added to 100 ml of hydrolysed porridge and incubated at 30°C for 24 h. 2.9. Preparation of munkoyo beverage using single organ&m starter cultures L. confusus starter culture (4 ml) was added, separately, to 400 ml of hydrolysed porridge to obtain an initial concentration of approximately 10 6 cfu/ml. S. cerevisiae starter culture (40 ml) was added to 400 ml of hydrolysed porridge to obtain an initial concentration of approximately 106 cfu/ml. 2.10. Effect of yeast to Lactobacillus ratio on the fermentation L. confusus and S. cerevisiae starter cultures were added to 400 ml of hydrolysed porridge to give the following approximate initial concentrations: (i) L. confusus, l 0 6 cfu/ml and S. cerevisiae, 102 cfu/ml; (ii) L. confusus, 106 cfu/ml and S. cerevisiae, 103 cfu/ml; and (iii) L. confusus, 106 cfu/ml and S. cerevisiae, 104 cfu/ml. The beverages were incubated at 30°C for up to 48 h. The aroma was evaluated by smelling them at each sampling time. Seven sterile control flasks, each containing sterile hydrolysed porridge, were also incubated. 2.1 I. Substrate utilisation during mixed culture fermentation Triplicate Erlenmeyer flasks, each containing 400 ml sterile Rhynchosia root extract-maize mixture, were inoculated with L. confusus starter culture to obtain a p p r o x i m a t e l y 10 6 cfu/ml and with S. cerevisiae starter culture, 10 2 cfu/ml. Mixtures were incubated at 30°C for 48 h and samples were removed from the flasks at intervals for enumeration of bacteria and yeasts and the assay of fermentation substrates and products. Seven sterile control flasks, each containing 25 ml sterile Rhynchosia root extract-maize mixture, were also incubated. 2.12. Sampling Duplicate samples (15 ml each) were removed from inoculated mixtures at 0, 1, 2, 3, 4, 6, 12, 18, 24, 30, 36, 42 and 48 h of incubation. Uninoculated flasks were 254 R.M. Zulu et al. / Int. J. Food Microbiology 34 (1997) 249-258 removed at 0, 6, 12, 18, 24, 36 and 48 h of incubation. One sample was frozen immediately for analysis of sugars, acids and alcohol, and the duplicate sample was used for microbiological, pH and preliminary sensoric evaluation by smelling. 3. Results 3.1. Amylolysis in sterile maize meal mixture by Rhynchosia sp. root extract The liquefaction was essentially complete in 1 h at 45°C, yielding primarily maltose (80% (w/w) of total sugars), maltotriose (17%) and glucose (3%) (Fig. 1). There was little or no further change in the sterile mixture during incubation at 30°C for 47 h (Fig. 1). The total sugars produced represented 75% (w/w) of the total starch in the Rhynchosia root extract-maize meal mixture. Mulkay et al. (1985) reported the presence of a-amylase and fl-amylase in root extracts of Rhynchosia insignis and Eminia holubii and both produced hydrolysates in which glucose concentrations were 6 5 - 7 5 % (w/w) of the maltose concentration. Maltotriose concentration was 23% (w/w) of maltose concentration in the R. insignis hydrolysate but undetectable in E. holubii hydrolysate. The hydrolysate prepared here contained very little glucose ( < 5% (w/w) of maltose concentration) and the concentration did not change substantially during the 1 h hydrolysis. It is not known if these differences are due to the use of different plant species (R. heterophylla) or to the use of different hydrolysis conditions. /( // 120 100 E E "-" Z 80 0 I-,< 60 n, I.-Z 40 Z 0 20 ILl ¢ O.O-O-O"--.-O-t~> 0 I t I I 0 0.2 0.4 0.6 I r 0.8 1.0 // I I I 6 24 48 TIME (h) Fig. 1. Hydrolysisof starch in maize meal porridge by Rhynchosia root extract at 45°C for 1 h, followed by 30°C for 48 h. &, maltose; T, maltotriose; ~, glucose. Values are means of three replicate hydrolyses and bars indicate ranges of determinations. 255 R.M. Zulu et al. / Int. J. Food Microbiology 34 (1997) 249-258 7 6 9 ! 8.5 I S ° L ~ o o14O t12o , ~ /~o °Io | ~ =.o 1 '°° =< 3 o 6.0 ~ ~ ~ i r : f i 0 6 12 18 24 30 36 42 ! - 8 48 TIME (h) Fig. 2. Growth of L. confusus LZI in Rhynchosia root extract-maize meal mixture at 30°C. 0 , L. confusus; [], pH value; i , maltose; ~?, maltotriose; O, ethanol; II, lactic acid. Concentrations of glucose and acetic acid were less than 5 and 2 mmol/1, respectively. 3.2. Fermentation of Rhynchosia root extract-maize meal mixture by single organisms Both L. confusus and S. cerevisiae grew well in Rhynchosia root extract-maize meal mixture on their own (Figs. 2 and 3). The yeast converted all (110 mmol/l) of the maltose to ethanol (460 mmol/1) but did not utilise maltotriose (Fig. 3). In the Lactobacillus culture there was an initial decrease in the maltose concentration followed by an apparent increase, presumably due to continuing amylolyic activity (Fig. 2). Maltotriose concentration showed little change and evidently maltotriose was not used by the bacterium. Presumably, the glucose was utilised but the concentrations were at all times below 5 mmol/1, the detection limits of the assay. The Lactobacillus produced lactic acid, which continued to increase after growth of the bacterium had ceased, and a little ethanol, but no acetic acid was detected ( < 2 mmol/1). The pH of the mixture with Lactobacillus fell to a final value of 3.5. The pH also fell rapidly in the yeast culture but only to 4.0. It is not known whether the decline in pH in the yeast culture was due to excretion of acids and/or the consumption of cations, such as ammonium. A faint 'munkoyo' aroma seemed to develop in the beverage inoculated with L. confusus starter culture alone but 'Munkoyo' aroma was not perceived in the beverage inoculated with S. cerevisiae alone (Fig. 5). Instead, a fruity-alcoholic aroma was apparent which persisted throughout the fermentation period. The uninoculated sterile beverage had an aroma of cooked maize meal-Rhynchosia root extract which was unchanged during the incubation period. 256 R.M. Zulu et al. / Int. J. Food Microbiology 34 (I997) 249 258 3.3. Effect of yeast to Lactobacillus ratio on the fermentation As expected, the larger the initial concentration of yeast the more rapid was the consumption of maltose and the production of ethanol (Fig. 4). All the beverages had the same initial concentration of Lactobacillus and the rate of production of lactic acid and decline in pH value was very similar in all the cultures, irrespective of the initial ratio of yeast to lactic acid bacteria. Thus, there was no evidence that the presence of the yeast had any effect on the growth of the lactic acid bacterium. Similarly, the pattern of growth, maltose consumption, and ethanol production by the yeast in the mixed cultures is indistinguishable from that in axenic yeast culture (Fig. 3). However, the final pH values of the mixed cultures were slightly lower (3.3) than in the axenic Lactobacillus fermentations (3.5) and it is possible that this was a consequence of combined acid production by both organisms. All the beverages inoculated with mixed inocula of L. confusus and S. cerevisiae developed a typical 'munkoyo' aroma. The higher the initial concentration of S. cerevisiae, the faster was the rate at which the 'munkoyo' aroma developed. The 'munkoyo' aroma persisted throughout the incubation period and was similar in all the beverages. 3.4. Substrate utilisation during mixed culture fermentation The concentration of maltotriose (18 mmol/1) remained almost constant throughout the fermentation period suggesting that it was not utilized by the microorgan8 r 1 500 E7.5 400 (3 300 < 5 ~> 7.0 200 -.-.-....A~ D - - .~ 6.5 D- - rn [] =o [] 100 8 0 6.0 0 6 12 18 24 30 36 42 48 TIME (h) Fig. 3. Growth of S. cerevisiae YZ20 in Rhynchosia root extract-maize meal mixture at 30°C. ©, S. cerevisiae; [3, pH value; A, maltose; T, maltotriose; O, ethanol. Concentrations of glucose and acetic acid were less than 5 and 2 mmol/l, respectively. 257 R.M. Zulu et al. / Int. J. Food Microbiology 34 (1997) 249 258 "~" o 500 7.0 400 6.0 300 5.0 LU E E Z 0 .-I In," I,Z I.IJ 200 z 0 lO0 4.0 "I- • D~D----~_.[D .,.~.A.. .... [] 3.0 / ""---cA , j i " . . / .- . . ~ z . . _ L j . - - - . ~ 2.0 ] I I I 0 6 12 18 I I I I I 24 30 36 42 48 T I M E (h) Fig. 4. Effect of ratio of S. cerevisiae YZ20 to L. confusus LZ1 on fermentation of Rhynchosia root extract-maize meal mixtures. - - - , initial S. cerevisiae, 8.9 x 104 cfu/ml; L. confusus, 2.0 x 106 cfu/ml; - - -, S. cerevisiae, 2.7 x 103 cfu/ml, L. confusus, 3.6 x l06 cfu/ml; .... S. cerevisiae, 6.3 x l02 cfu/ml; L. confusus 3.6 x 106 cfu/ml; A, maltose; e , ethanol; II, mean lactic acid in all 3 fermentations; [~, mean pH value of 3 fermentations. isms (Fig. 5). Glucose (6 mmol/1) was detected only in samples before fermentation and, thereafter, glucose was not detected, presumably due to utilization by the microorganisms. 3.5. Preparation of munkoyo beverage using starter cultures S. cerevisiae alone was not suitable as a starter culture for the preparation of munkoyo beverage because it failed to elicit the typical munkoyo flavoured beverage. L. confusus alone did generate a faint munkoyo aroma and a beverage prepared with L. confusus alone might be acceptable to some consumers. The development of a full 'munkoyo' aroma required the presence of the Lactobacillus and the yeast. Rhynchosia root extract-maize meal mixtures fermented for 48 h with mixed cultures of L. confusus and S. cerevisiae produced beverages with a pH of about 3.5 and a typical 'munkoyo' aroma but contained an alcohol content (ca. 400 mmol/1) greater than the required level ( < 100 mmol/1). To obtain a munkoyo beverage with the desired low alcohol concentration, it is evident that the ratio of yeast concentration to Lactobacillus concentration (cfu/ml) in the starter culture should be 1:1000 or less and the beverage should be fermented for 24 h only (Figs. 4 and 5). 258 R.M. Zulu et al. / Int. J. Food Microbiology 34 (1997) 249-258 AIO 400 o (J o oi 8 3oo E E 0 0 200 <~ z Z 0 0 _J <~ 0 n, 6 n, I-2: 11,1 100 0 Z 4 o (.I :Z 2 i i i i i u i n i 0 6 12 18 24 30 36 42 48 TIME (h) Fig. 5. Fermentation of Rhynchosia root extract-maize meal mixture by S. cerevisiae YZ20 and L. confusus LZ1 at 30°C. O, Yeast concentration; R, bacterial concentration; A, maltose; V, maltotriose; O, ethanol; m, lactic acid. Values are means of three replicate fermentations and bars indicate ranges of determinations. Concentrations of glucose and acetic acid were less than 5 and 2 mmol/1, respectively. Acknowledgements RMZ thanks the National Council for Scientific Research, Zambia and the EEC for financial support. References Graffham, A.J. (1994) Tempe fermentation as a Processing Option for African Cowpea. PhD Thesis, The University of Reading, Reading. Holzapfel, W.H. (1989) Industrialization of mageu fermentation in South Africa. In: K.H. Steinkraus (editor), Industrialization of Indigenous Fermented Foods, Marcel Dekker, New York, pp. 285-328. Mulkay, P., Delaude, C. and Huls, R. (1985) Le pouvoir amylotyque de Rhynchosia insignis (Fabaceae). Bull. Soc. Roy. Sci. Liege 54, 187 193. Nuraida, L., Wacher, M.C. and Owens, J.D. (1995) Microbiology of pozol, a Mexican fermented maize dough. World J. Microbiol. Biotech. 11, 567 571. Pauwels, L., Mulkay, P., Ngoy, K. and Delaude, C. (1992) Eminia, Rhynchosia et Vigna (Fabac6es) complexes amylolytiques employ6s dans la r6gion Zamb6sienne pour la fabrication de la bi6re 'Munkoyo'. Belg. J. Bot. 125, 41-60. Simwamba, C.G. and Elahi, M. (1986) Studies on the nutrient composition of Rhynchosia venulosa (munkoyo roots) and physicochemical changes in munkoyo roots and maize porridge mixture during preparation of munkoyo beverage. J. Agric. Food Chem. 34, 573-575. Steinkraus, K.H. (editor) (1996) Handbook of Indigenous Fermented Foods, 2"a edn., Marcel Dekker, New York, pp. 437 439. Zulu, R.M. (1993) Structure Elucidation of Flavonoid Compounds of Munkoyo Roots Extract (Rhynchosia species) and Studies on Munkoyo Beverage Fermentation. PhD Thesis, The University of Reading, Reading.