the growth and interaction of yeasts and lactic acid bacteria isolated from zimbabwean naturally...

Ž .International Journal of Food Microbiology 68 2001 21–32www.elsevier.comrlocaterijfoodmicro

The growth and interaction of yeasts and lactic acid bacteriaisolated from Zimbabwean naturally fermented milk in UHT milk

Tendekayi H. Gadagaa,b,), Anthony N. Mutukumiraa, Judith A. Narvhusba Institute of Food, Nutrition and Family Sciences, UniÕersity of Zimbabwe, P.O. Box MP 167, Mt. Pleasant, Harare, Zimbabwe

b ˚Department of Food Science, Agricultural UniÕersity of Norway, P.O. Box 5036, N-1432, As, Norway

Received 7 August 2000; received in revised form 17 December 2000; accepted 6 January 2001

Abstract

Nine yeast and four lactic acid bacterial strains, previously isolated from Zimbabwean traditionally fermented milk, wereŽ .inoculated into ultra-high temperature treated UHT milk in both single and yeast–lactic acid bacteria co-culture. The lactic

Ž .acid bacteria LAB strains consisted ofLactococcus lactis subsp.lactis biovar. diacetylactis C1, L. lactis subsp.lactisLc39, L. lactis subsp. lactis Lc261 and Lactobacillus paracasei subsp. paracasei Lb11. The yeast strains used wereCandida kefyr 23, C. lipolytica 57, C. lusitaniae 63, C. lusitaniae 68, C. tropicalis 78, Saccharomyces cereÕisiae 71, S.dairenensis 32, C. colliculosa 41 and Dekkera bruxellensis 43. After 48-h fermentation at 258C, the samples were analysed

Ž .for pH, viable yeast and bacterial counts, organic acids, volatile organic compounds VOC and carbon dioxide. TheLactococcus strains reduced the pH from about 6.6 to between 4.0 and 4.2, whileLb. paracasei subsp. paracasei Lb11reduced the pH to about 5.4. Most of the yeasts, however, did not affect the final pH of the milk except forC. kefyr 23,which reduced the pH from 6.6 to 5.8. All theLactococcus strains grew two log cycles during the 48-h fermentation period,while Lb. paracasei subsp. paracasei Lb11 grew about one log cycle.S. cereÕisiae 71, C. colliculosa 41 and D.bruxellensis 43 showed poor growth in the milk in both single and co-culture. The other species of yeast grew about two logcycles.Candida colliculosa 41, S. dairenensis 32 and D. bruxellensis 43 showed reduced viability when in co-culture withLb. paracasei subsp.paracasei Lb11. The samples in whichC. kefyr 23 was used were distinct and characterised by largeamounts of acetaldehyde, carbon dioxide and ethanol. However, in the samples whereS. dairenensis, C. colliculosa, D.bruxellensis, C. lusitaniae, C. tropicalis, C. lipolytica and S. cereÕisiae were used in co-culture, the final pH and metabolitecontent were mainly determined by the corresponding LAB strain. From the observations, it was concluded thatC. kefyr 23grows well in UHT milk and produces VOC that may be important to the flavour profile of the fermented milk.Enhancement of production of some flavour compounds such as acetaldehyde and malty compounds in some yeast–LABco-cultures was presumed to be indicative of interaction between the microorganisms.q2001 Elsevier Science B.V. Allrights reserved.

Keywords: Yeasts; Fermented milk; Flavour compounds; Organic acids

) Corresponding author. Institute of Food, Nutrition and FamilySciences, University of Zimbabwe, P.O. Box MP 167, Mt. Pleas-ant, Harare, Zimbabwe. Fax:q47-64-94-37-89.

Ž .E-mail address: [email protected] T.H. Gadaga .

1. Introduction

Ž .Yeasts, together with lactic acid bacteria LAB ,are part of the microbial flora of Zimbabwean natu-

0168-1605r01r$ - see front matterq2001 Elsevier Science B.V. All rights reserved.Ž .PII: S0168-1605 01 00466-4

( )T.H. Gadaga et al.r International Journal of Food Microbiology 68 2001 21–3222

Žrally fermented milk Feresu and Muzondo, 1990;Mutukumira, 1995; Mutukumira, 1996; Gadaga et

.al., 2000 . Earlier work showed thatSaccharomycescereÕisiae, Candida lusitaniae, C. colliculosa and S.dairenensis are some of the predominant yeast

Ž .species isolated from amasi Gadaga et al., 2000 .Yeasts affect the quality of fermented milk throughthe production of flavour compounds and other

Ž .metabolites Jakobsen and Narvhus, 1996 . Lactosefermentation and assimilation, lactate assimilation,lipolysis and proteolysis are the important reactionsof yeasts which are responsible for the diverse flavour

Žcompounds reported in many dairy products Roostita.and Fleet, 1996 .

Several microbial interactions involving yeastshave been suggested in fermented products such asblue cheese, white mould cheese, bacterial surface

Žripened cheeses, kefir and koumiss Subramanianand Shankar, 1985; Fleet, 1990; Jakobsen and

. Ž .Narvhus, 1996 . Marshall 1987 reviewed possibleinteraction between yeasts and lactic acid bacteria infermented milk, but the mechanisms of these interac-tions have not been elucidated. Interaction betweenLactobacillus hilgardii and S. florentinus isolatedfrom sugary kefir grains has also been reported,where the yeast stimulated the LAB through produc-tion of carbon dioxide, pyruvate, propionate and

Ž .succinate Leroi and Pidoux, 1993a . In addition,some LAB release galactose into the medium as a

Žby-product of lactose metabolism Marshall, 1987;Davidson and Hillier, 1995; Marshall and Tamime,

.1997 , which may be used by galactose-assimilating,but lactose-negative, yeasts. Proteolytic yeasts suchas Yarrowia lipolytica and C. catenulata grow inmilk and produce free amino acids such as leucine,phenylalanine, lysine, arginine, glutamic acid and

Ž .valine Roostita and Fleet, 1996 , which can be asource of metabolisable substrate for other microor-ganisms, resulting in the production of secondarymetabolites including flavour compounds. Release offree amino acids could also promote the growth ofLAB with a poor proteolytic system.

The high yeast counts in Zimbabwean traditionalŽ y1.fermented milk ca. 6.9 log cfu g suggest that10

the yeasts have a mechanism for growth in the milkŽ .Mutukumira, 1996; Gadaga et al., 2000 . The aim ofthis study was to assess the ability of yeasts isolatedfrom Zimbabwean naturally fermented milk to grow

Ž .in ultra-high temperature UHT treated milk. Possi-ble interactions between the yeasts and LAB werealso investigated by studying final populations in themilk and by comparing their ability to producevolatile compounds, organic acids and carbon diox-ide in pure and co-culture.

2. Materials and methods

2.1. Storage and transportation of yeast isolates

Yeasts strains isolated from Zimbabwean fer-Ž .mented milk Gadaga et al., 2000 were stored on

Ž . Žmalt extract agar MEA Merck, Damstadt, Ger-.many slants at 48C and were transported to Norway

in an insulated cooler bag with ice packs. The yeastswere further stored on MEA slants at 48C at theAgricultural University of Norway.

2.2. Selection of yeast strains

Nine yeast strains were selected from 44 isolates,based on their ability to produce relatively high

Ž .levels of volatile organic compounds VOC andŽcarbon dioxide in UHT milk Tine Norske Meierier,

.Oslo, Norway , after fermentation for 48 h at 258CŽ .results not shown . The following yeast strains wereselected:C. kefyr 23, S. cereÕisiae 71, C. lipolytica57, C. lusitaniae 68, C. tropicalis 78, C. lusitaniae63, C. colliculosa 41, S. dairenensis 32 andDekkerabruxellensis 43. All the cultures are at present storedin the culture collection of the Department of FoodScience, Agricultural University of Norway. Concen-trated cultures of the yeasts were then prepared.

2.3. Preparation of concentrated yeast cultures

Each actively growing culture of the selectedyeasts was inoculated separately into malt extract

Ž . Ž .broth Merck 200 ml in 250-ml screw cappedbottles and incubated for 72 h at 258C in a waterbath. The cultures were then centrifuged at 5000=g

Ž . Žin a refrigerated 4–108C centrifuge Sorval 5RB,.Du Pont Instruments, DE, USA . The pellet was

Ž .re-suspended in sterile reconstituted 10%, wrvŽ . Ž .skimmed milk 20 ml containing 10% vrv glyc-

erol and then stored aty808C.

( )T.H. Gadaga et al.r International Journal of Food Microbiology 68 2001 21–32 23

2.4. Preparation of concentrated lactic acid bacteriacultures

Concentrated pure cultures ofLactococcus lactissubsp. lactis biovar. diacetylactis C1, L. lactissubsp.lactis Lc39, L. lactis subsp.lactis Lc261 andLb. paracasei subsp. paracasei Lb11, previouslyisolated from Zimbabwean traditional fermentedmilk, were prepared. TheLactococcus and theLac-tobacillus strains were inoculated into M17 broth

Žand MRS broth Oxoid Unipath, Hampshire, Eng-. Ž .land 250 ml , respectively, and incubated at 308C

for 24 h. The cultures were centrifuged in a similarway to the yeast cultures and then re-suspended in

Ž .25 ml of reconstituted 10%, wrv skimmed milk.The cultures were stored aty808C and later used as

Ž .direct vat set DVS cultures for inoculating UHTmilk.

2.5. Inoculation of UHT milk with DVS cultures

The thawed concentrated cultures had viable LABand yeast counts ranging between 9.81–10.3 and7.6–8.3 log cfu mly1, respectively. A volume of10

the individual concentrated cultures was added to theŽ .UHT milk 40 ml , calculated to obtain a standard

initial inoculum of about 7.0 log cfu mly1 LAB10y1 Ž .and 5.0 log cfu ml yeast Tables 2 and 3 . A10

Ž .portion of each inoculated milk sample 10 ml wasaseptically pipetted into a headspace vial for carbondioxide analysis and incubated together with the restof the milk at 258C for 48 h. The viable microbialcounts, VOC, organic acids, carbon dioxide and pHwere determined in the incubated milk.

2.5.1. Viable microorganismsLactococcus and Lactobacillus strains were enu-

Ž . Ž .merated pour plate on M17 and MRS agar Oxoid ,respectively, and incubating at 308C for 48 h. Yeasts

Ž .were enumerated spread plate on yeast extract glu-Ž . Ž .cose chloramphenicol agar YGCA IDF, 1990 and

incubated at 258C for 3 days.

2.5.2. Chemical analysesŽpH was determined using a Radiometer model

.PHM92 pH meter with a combination glass elec-Žtrode and temperature probe Radiometer, Copen-

.hagen after calibrating with commercial buffersŽ . Ž .Merck pH 4 and 7 . Volatile organic compounds

Ž .VOC were determined by headspace gas chro-Ž .matography HS-GC and the organic acids were

determined using high performance liquid chro-Ž .matography HPLC according to the method of

Ž .Narvhus et al. 1990; 1998 . The VOC determinedwere acetaldehyde, ethanol, acetone, diacetyl, ace-toin, 3-methyl-butanal, 3-methyl-butanol, 2-methyl-butanal and 2-methyl-butanol. The following organicacids were determined: citric, orotic, pyruvic, suc-cinic, lactic, formic, uric, and propionic acids. Car-bon dioxide was determined using the method of

Ž .Narvhus et al. 1991 . All the determinations weredone in duplicate.

2.6. Statistical analysis

Average values of VOC, organic acids and carbondioxide data were analysed using principal compo-

Ž . Ž .nent analysis PCA with cross-validation using thew Ž wUnscrambler Programme Unscrambler 7.01,

.Camo, Trondheim, Norway . Each variable wasweighted by dividing with the standard deviation ofthat variable. Interpretation of the data was made byinspection of the scores and loadings plots. Singleand co-cultured samples were compared by perform-ing one-way analysis of variance using Minitab v

Ž .12.02 Minitab programme.

3. Results

The results are presented according to changes inmicrobial cell counts, pH, organic acids, VOC andproduction of carbon dioxide after 48-h fermentation.The values given are the mean from two independentexperimental runs.

3.1. Viable microorganisms

The Lactococcus strains grew approximately twolog cycles both in single and co-cultures. The finalnumbers of theLactococcus strains in the fermentedmilk were in the range 9.04–9.51 log cfu mly1.10

Lb. paracasei subsp.paracasei Lb11 grew approxi-mately one log cycle and the viable counts ranged

y1 Ž .between 8.43 and 8.84 log cfu ml Table 1 in10

both single and co-culture. The viable counts of thethree Lactococcus strains were not changed by co-inoculation with yeast, except for lower viable counts

()

T.H

.Gadaga

etal.r

InternationalJournalofF

oodM

icrobiology68

200121

–32

24

Table 1The viable LAB counts in UHT milk after 48-h fermentation with selected combinations of yeasts and LAB strains

y1Ž .LAB strain Initial inoculum Final LAB count Lactic acid bacteria counts after growth with yeasts log cfu ml10y1 y1Ž . Ž .log cfu ml log cfu ml10 10 C. kefyr C. lipolytica S. cereÕisiae C. lusitaniae C. tropicalis C. lusitaniae C. colliculosa S. dairenensis D. bruxellensis

23 57 71 68 78 63 41 32 43

L. lactis subsp. 7.45 9.3 9.30 9.42 9.40 9.39 9.29 9.51 9.42 9.35 9.44lactis biovar.diacetylactis C1L. lactis subsp. 7.98 9.35 9.31 9.04 9.26 9.34 9.21 9.41 9.33 9.24 9.19lactis Lc39

a a a a aL. lactis subsp. 7.19 9.06 9.38 9.22 9.39 9.31 9.25 9.25 9.19 9.34 9.22lactis Lc261

a a aLb. paracasei 7.93 8.68 8.81 8.64 8.68 8.84 8.70 8.53 8.59 8.74 8.45subsp.paracaseiLb11

a Ž .Final LAB counts in co-cultures were significantlyPF0.01 higher than the single culture. Results are from two replicate trials.


Table 2The viable yeast counts in UHT milk after 48-h fermentation with selected combinations of yeast and LAB strains

y1Yeast strain Initial yeast Final yeast counts in Final yeast counts log cfu ml with LABy1 y1Ž . Ž .inoculum cfu ml single culture cfu ml L. lactis subsp. L. lactis L. lactis Lb. paracasei

lactis biovar. subsp. lactis subsp. lactis subsp.paracaseidiacetylactis C1 Lc39 Lc261 Lb11

C. kefyr 23 5.90 7.45 7.27 7.32 7.86 7.19C. lipolytica 57 5.30 6.90 6.41 7.03 6.33 6.66

aS. cereÕisiae 71 5.60 6.45 6.08 6.52 5.78 6.66C. lusitaniae 68 5.00 7.52 7.16 7.55 7.94 6.90C. tropicalis 78 5.48 7.21 7.19 7.20 6.58 7.02C. lusitaniae 63 5.85 7.65 7.32 7.28 7.29 6.83

aC. colliculosa 41 5.78 6.86 6.56 6.74 6.65 6.00aS. dairenensis 32 5.70 7.21 6.48 6.40 6.78 5.78aD. bruxellensis 43 6.42 6.30 6.57 6.45 6.00 5.62

a Ž .The final yeast counts were significantlyPF0.01 lower in co-culture than in single culture. The results are from two replicate trials.

of L. lactis subsp. lactis Lc39 when in co-culturewith C. lipolytica 57. The final viable counts forLc261 were comparatively higher in co-culture withC. kefyr 23, S. cereÕisiae 71, andS. dairenensis 32than in single culture.Lb. paracasei subsp.paraca-sei Lb11 had higher viable counts in co-culture withC. kefyr 23, C. lusitaniae 68 andS. dairenensis 32Ž .Table 1 .

The ability of the selected yeast cultures to growin milk was variable. The final populations were inthe range of 5.45–7.65 log cfu mly1. S. cereÕisiae10

71, C. colliculosa 41 and D. bruxellensis 43 insingle culture showed poor ability to grow in the

Ž .UHT milk Table 2 . C. kefyr 23 grew to finalpopulations of 7.19–7.86 log cfu mly1. C. collicu-10

losa 41, S. dairenensis 32 and D. bruxellensis 43when grown in co-culture with Lb11 had a lowerfinal population than in the original inoculum, sug-gesting that their viability may have been reduced at

Ž .the end of the fermentation Table 2 .

3.2. pH

L. lactis subsp.lactis biovar. diacetylactis C1, L.lactis subsp.lactis Lc39 and L. lactis subsp.lactisLc261, in single and co-culture, reduced the pH from6.6 to levels ranging between 4.0 and 4.2.Lb.paracasei subsp.paracasei Lb11 lowered the pH to

Žabout 5.4 after the 48-h fermentation period Table.3 , and consequently, did not coagulate the milk.

However, the yeast strains in single culture did notcause any notable changes in the pH of the UHTmilk except C. kefyr 23, which reduced the pH to

Ž .5.8 Table 3 .

3.3. Organic acids, VOC and carbon dioxide

Data for organic acids, VOC and carbon dioxideproduced by the different lactic acid bacteria and

Ž .yeast cultures, were analysed using PCA Fig. 1 .The first two principal components accounted for

Table 3The pH of UHT milk after 48-h fermentation with selected combinations of yeast and LAB strains

aLAB strain pH in pure LAB culture pH in co-culture with yeasts

L. lactis subsp.lactis biovar. diacetylactis C1 4.2"0.09 4.18–4.23L. lactis subsp.lactis Lc39 4.1"0.08 4.05–4.16L. lactis subsp.lactis Lc261 4.2"0.03 4.04–4.13Lb. paracasei subsp.paracasei Lb11 5.4"0.04 5.08–5.43

aThe range of pH of samples fermented with the LAB strains in co-culture withC. kefyr 23, C. lipolytica 57, C. lusitaniae 63, C.lusitaniae 68, C. tropicalis 78, S. cereÕisiae 71, S. dairenensis 32, C. colliculosa 41, and D. bruxellensis 43. The results are from tworeplicate trials.


41% and 25%, respectively, of the variation in thedata.

The samples fermented by the various culturesŽ .could be clustered into five groups A–E as shown

Ž .on the PCA scores and loadings bi-plot Fig. 1 .Group A comprised all the pure yeast cultures exceptC. kefyr 23, and were grouped together on the PCAplot by the organic acids: succinic, orotic and citricacids. These compounds were in high concentrationsat the end of the fermentation, similar to those in theoriginal milk. The initial concentrations of succinic,orotic and citric acid in the UHT milk were 565, 80and 1925 mg kgy1, respectively. However,C. lipoly-tica 57 andC. tropicalis 78 assimilated succinic acidŽ .Fig. 2a . C. kefyr 23 was placed away from theother yeasts on the PCA bi-plot due to the largeamounts of carbon dioxide, aldehydes and alcoholsproduced.C. kefyr 23 also showed strong utilisationof succinic acid. It was also noted that none of theyeast strains produced detectable levels of methylaldehydes except forC. kefyr 23, which produced

y1 Ž0.5 mg kg of 2-methyl-propanal results not.shown . Levels of 2-methyl-1-propanol, 2-methyl-1-

butanol, and 3-methyl-1-butanol were also high inthis culture compared to the other starter cultures and

Ž .had high loadings on both PCs Fig. 1 .Group B comprised all of the samples fermented

by L. lactis subsp.lactis Lc261 and Lb. paracaseisubsp. paracasei Lb11 in pure and co-culture, ex-

cept for the Lc261r23 and Lb11r23 cultures. L.lactis subsp.lactis Lc261 andLb. paracasei subsp.paracasei Lb11 also did not utilise succinic acid andproduced low levels of carbon dioxide, acetaldehydeand the alcohols. However,L. lactis subsp. lactisLc261 produced high levels of lactic acid of up to0.8 g 100 gy1 in co-culture with C. tropicalis 78Ž .results not shown . The levels of orotic acid wereslightly reduced from 80 to 73 mg kgy1 when Lb.paracasei subsp. paracasei Lb11 was co-culturedwith S. cereÕisiae 71 and D. bruxellensis 43, eventhough the individual LAB and yeast strains did not

Ž .utilise this compound results not shown .Group C contained samples fermented with the

LAB culture L. lactis subsp. lactis Lc39 and itsco-cultures with yeasts.L. lactis subsp.lactis Lc39

Žutilised succinic acid to below detection levels Fig..2a . This property was not affected by addition of

yeasts. This group was further characterised by 2-Žmethyl-propanal and 3-methyl-butanal results not

.shown . Negligible amounts of 2-methyl-butanalŽ .were produced results not shown . The correspond-

ing alcohols were detected in high quantities, espe-Ž .cially 3-methyl-1-butanol Fig. 3c , while 2-methyl-

1-propanol that was produced in co-culture withyeasts, exceptC. kefyr 23, was in the range 2.4–3.1mg kgy1. 2-Methyl-1-butanol was produced in therange 0.3–0.5 mg kgy1 in co-culture, and this was

Ž y1.slightly less than in single culture 0.7 mg kg .L.

Fig. 1. The scores and loadings bi-plot for volatile organic compounds, organic acids and CO produced byL. lactis subsp.lactis biovar.2

diacetylactis C1, L. lactis subsp.lactis Lc39, L. lactis subsp.lactis Lc261, andLb. paracasei subsp.paracasei Lb11 in single culture andin co-culture withC. kefyr 23, C. lipolytica 57, C. lusitaniae 63, C. lusitaniae 68, C. tropicalis 78, S. cereÕisiae 71, S. dairenensis 32, C.colliculosa 41, andD. bruxellensis 43 in UHT milk after 48-h fermentation. Group B includes all cultures fermented with Lc261 and Lb11Ž . Ž .except Lc261r23 and Lb11r23 , while group C includes all cultures fermented with Lc39 except Lc39r23 .


Ž . Ž . Ž .Fig. 2. The levels of a succinic acid, b pyruvic acid, c formicŽ .acid, and d carbon dioxide in UHT milk after 48-h fermentation

Ž .with yeasts aloneW , yeast with L. lactis subsp.lactis biovar.Ž .diacetylactis C1 Y , yeasts with L. lactis subsp. lactis Lc39

Ž . Ž ., yeasts with L. lactis subsp. lactis Lc261 I and yeastsŽ .with Lb. paracasei subsp.paracasei Lb11 B . The yeast strains

used wereC. kefyr 23, C. lipolytica 57, C. lusitaniae 63, C.lusitaniae 68, C. tropicalis 78, S. cereÕisiae 71, S. dairenensis32, C. colliculosa 41 and D. bruxellensis 43. The LAB were also

Ž .inoculated in single culture LAB alone . The results are from tworeplicate trials. The bars represent standard deviations.

lactis subsp.lactis Lc39 produced up to 0.77 g 100gy1 lactic acid in both single and co-culture.

Group D consisted of samples fermented with C1.These samples were distinctly different from theother LAB cultures due to the higher levels of lacticacid, diacetyl, acetoin, 2-butanone and the maltycompounds.L. lactis subsp.lactis biovar. diacety-lactis C1 in single culture produced up to 0.9 g 100gy1 lactic acid, 1.3 mg kgy1 diacetyl and 86 mg

y1 Ž .kg acetoin results not shown . When C1 wascultured in combination with yeasts, the levels ofdiacetyl and acetoin were reduced, although the ef-fect was less marked in the case of acetoin. Reduc-tion of diacetyl and acetoin was most notable for theco-cultures containingC. tropicalis 78. The co-cul-tures of C1 withS. cereÕisiae 71, S. dairenensis 32,C. colliculosa 41 and D. bruxellensis 43 producedsimilar levels of 2-methyl-1-propanol, 2-methyl-1-

Ž .butanol results not shown , 3-methyl-1-butanol, ac-Ž .etaldehyde and ethanol Fig. 3a–c .

All of the LAB and yeast cultures produced car-bon dioxide as they grew in the UHT milk, althoughthis observation was more pronounced for the citrate

Ž .positive C1 strain Fig. 2d . Of the yeast cultures,C.kefyr 23 produced the highest levels of carbon diox-

Ž y1.ide 3000–4000 mg kg . The samples containingŽC. kefyr 23 were very different from the others Fig.

.1 . Although they were not closely clustered togetheron the PCA bi-plot, they were grouped together byhigh levels of carbon dioxide, acetaldehyde and thealcohols compared to the other samples. In thesesamples, ethanol had high positive loadings on thefirst PC and was produced in the range 0.5–0.8 g

y1 Ž .100 g Fig. 3b . The co-cultures ofS. cereÕisiae71, S. dairenensis 32, C. colliculosa 41 and D.bruxellensis 43 produced more ethanol than the sin-

Ž .gle cultures Fig. 3b . The C1r23 co-culture wasfurther characterised by the malty compounds 2-methyl-butanal and 2-methyl-propanal, which wereproduced both by the yeast and LAB.Lb. paracaseisubsp.paracasei Lb11 andC. kefyr 23 in co-culturealso produced high levels of acetone of up to 12.3mg kgy1. The other yeast–LAB co-cultures hadacetone levels less than 2 mg kgy1, but C. tropicalis78 seemed to considerably reduce the amount of

Ž .acetone in the products results not shown .ŽAll the LAB strains produced acetaldehyde Fig.

.3a . The highest amounts of acetaldehyde were pro-


Ž . Ž . Ž .Fig. 3. The levels of a acetaldehyde, b ethanol and c 3-methyl-1-butanol in UHT milk after 48-h fermentation with yeasts

Ž .alone W , yeast withL. lactis subsp.lactis biovar. diacetylactisŽ . Ž .C1 Y , yeasts withL. lactis subsp.lactis Lc39 , yeasts with

Ž .L. lactis subsp.lactis Lc261 I and yeasts withLb. paracaseiŽ .subsp.paracasei Lb11 B . The yeast strains used wereC. kefyr

23, C. lipolytica 57, C. lusitaniae 63, C. lusitaniae 68, C.tropicalis 78, S. cereÕisiae 71, S. dairenensis 32, C. colliculosa41, and D. bruxellensis 43. The LAB were also inoculated in

Ž .single culture LAB alone . The results are from two replicatetrials. The bars represent standard deviations.

duced whenC. kefyr 23 was cultured together withŽL. lactis subsp. lactis biovar. diacetylactis C1 72

y1.mg kg andLb. paracasei subsp.paracasei Lb11Ž y1. Ž .37 mg kg Fig. 3a . Additionally,C. kefyr 23produced the highest levels of ethanol when growntogether with Lb. paracasei subsp.paracasei Lb11Ž y1. Ž .0.8 g 100 g Fig. 3b . However,Lb. paracaseisubsp. paracasei Lb11 in single culture produced

Ž y1.very little acetaldehyde 2.5 mg kg and ethanolŽ y1.7.2 mg kg , whileC. kefyr 23 produced 0.7 g100 gy1 ethanol.

All the LAB strains also produced detectableŽ .levels of pyruvic acid in the samples Fig. 2b .

Among the yeasts, onlyC. kefyr 23 produced de-tectable levels of pyruvic acid.C. kefyr 23 producedthe highest amounts of pyruvic acid in single cultureŽ y1.33 mg kg , while its co-culture with Lc39 also

Ž .had high values Fig. 2b . These amounts of pyru-vate produced by the LAB were comparatively lowerwhen C. tropicalis 78 was co-inoculated.L. lactissubsp. lactis Lc261 produced the lowest levels ofpyruvate in both single and co-culture. Formic acidwas only detected in theLactococcus cultures in the

y1 Ž .range 25–75 mg kg Fig. 2c . All LAB cultures,both in pure and co-culture, producedDL-pyro-

Ž y1.glutamic acid 3–11 mg kg , but none of theŽ .yeasts produced this metabolite results not shown .

Propionic acid was not detected in any of the sam-ples.

4. Discussion

The occurrence of yeasts together with LAB andcoliforms in naturally fermented milk has led tosuggestions of possible interactions between thesegroups of microorganisms. Understanding the rolesof the yeasts in the milk will help in the quest todevelop cultured milks that have similar character-istics to traditional fermented milk.

The LAB strains were found to be responsible foracidifying the milk while the yeasts caused littlechange in the pH. However, there was a slightdecrease in pH in the milk cultured withC. kefyr 23.This may partly be attributed to production of acidic

Ž .compounds such as acetic acid not determined orproteolysis and lipolysis.KluyÕeromyces marxianusŽ .perfect state ofC. kefyr excretes proteases and

Žlipases that hydrolyse milk protein and fat Roostita.and Fleet, 1996 . However, the proteolytic and

lipolytic properties of C. kefyr 23 have not beenstudied.

The lactococci grew from about 7 to 9 log cfu10

mly1 in the UHT milk, while the yeasts could reachmaximum populations of 7.65 log cfu mly1 from10


an initial inoculum level of 5 log cfu mly1. These10

LAB counts are comparable to those encounteredŽwith other reference strains and dairy starters Teuber,

.1995 , and correspond to the lactic acid and pHvalues shown in this study. However, Lb11 showedless potential to grow in the UHT milk. This trait hasbeen reported for some wild strains ofLactobacillus

Žisolated from naturally fermented milk Fekadu,.1994; Sserunjogi, 1999 . However, the higher popu-

lations of Lb. paracasei subsp. paracasei Lb11recorded in co-culture withC. kefyr 23 suggest thatthe yeast stimulated growth of the LAB. The yeastcould achieve this by providing essential metabolitessuch as pyruvate, amino acids and vitamins.

The growth response of the yeasts can be com-Ž .pared to results reported by Roostita and Fleet 1996 .

In that study, yeasts were able to grow to maximumpopulations of 107–108 cfu mly1 when grown inUHT milk. The growth of C. kefyr 23 can beexplained by the strain’s ability to utilise lactose.S.cereÕisiae 71, C. colliculosa 41 and D. bruxellensis43 showed poor growth in the milk. ForS. cere-Õisiae 71, this can be explained by the fact that thisspecies does not metabolise lactose because it lacks atransport mechanism for lactose into the yeast cellŽ .Walker, 1998 . Some studies, however, have sug-gested that some strains ofLactococcus and Lacto-bacillus metabolise the glucose, but not the galactosemoeity of lactose, which is then secreted into the

Žmedium Marshall, 1987; Davidson and Hillier, 1995;.Montanari et al., 1996; Marshall and Tamime, 1997 ,

which could partly explain the growth of the non-lactose fermenting yeasts in the fermented milk.However, other researchers have connected thegrowth of S. cereÕisiae in cultured milk with its

Žability to metabolise lactic acid Fleet, 1992; Sarais.et al., 1996 .C. colliculosa 41, S. dairenensis 32

and D. bruxellensis 43 had lower viable countswhen co-cultured with Lb11, which could suggestthat there was an antagonistic effect of the LAB onthe yeast, or insufficient metabolites. For example,these yeast strains assimilate lactate but Lb11 onlyproduced low amounts of lactic acid. However, abil-ity to assimilate lactate does not explain why theywere able to grow in single culture. Some othermechanisms such as utilisation of milk proteins andfats may also be important.C. lipolytica 57 seemedto slightly suppress the growth of Lc39 when grown

in co-culture. This could probably be due to the factthat C. lipolytica is strongly lipolytic and may pro-duce free fatty acids that are inhibitory to the lactic

Žacid bacteria Anders and Jago, 1970; Broome et al.,.1979; Venugopal, 2000 . This relationship, however,

needs to be investigated further.VOC, organic acids and carbon dioxide are im-

portant for the sensory properties of fermented milk.In addition, production of metabolites, or use ofsubstrates can assist in establishing the basis forgrowth of test strains in milk. Most of the yeaststrains did not reduce succinic, orotic and citric acidsin the UHT milk, which could suggest that they donot use these compounds as carbon sources. How-ever,C. kefyr 23, C. lipolytica 57 andC. tropicalis78 assimilated succinic acid, which is indicative ofoxidative metabolism of carbohydrates through the

Ž .citric acid cycle Walker, 1998 . Roostita and FleetŽ .1996 also reported that a strain ofK. marxianusthey used in their studies showed partial utilisationof citric acid, formic and succinic acids.

Only the alcoholic malty compounds were de-tected in milk cultured withC. kefyr 23. It ispossible that theC. kefyr quickly reduced the alde-

Žhyde malty compounds to the alcohols Stam et al.,.1998 . L. lactis subsp. lactis Lc39 also produced

high levels of malty compounds. At the end of thefermentation period, the alcohols were at higherlevels than the aldehydes. The malty compounds

Žhave low minimum taste thresholds Peppard and.Halsey, 1981; Narvhus et al., 1998 , such that the

levels of the malty compounds produced byL. lactissubsp. lactis Lc39 may result in the production ofunacceptable cultured milk. In the presence of yeasts,the levels of 2-methyl-1-propanol and 3-methyl-1-butanol produced byL. lactis subsp. lactis Lc39were higher, showing that, in the presence of yeasts,the reduction of the aldehydes to the alcohols isenhanced. The levels of malty compounds producedby L. lactis subsp. lactis biovar. diacetylactis C1were much lower than those produced byL. lactissubsp. lactis Lc39. L. lactis subsp. lactis biovar.diacetylactis C1 produces fermented milk of accept-

Žable quality Mutukumira, 1996; Narvhus et al.,.1998 , despite the presence of low levels of malty

compounds. A complex mixture of flavour com-pounds, however, may determine the overall accep-tance of the fermented milk.


The high levels of acetaldehyde produced byC.kefyr 23 in co-culture with all LAB, especiallyL.lactis subsp.lactis biovar. diacetylactis C1 and Lb.paracasei subsp. paracasei Lb11 can be comparedto those produced by some types of yoghurt culturesŽ .Tamime and Robinson, 1999 . However, acetalde-hyde levels of up to 72 mg kgy1 produced by theco-culture ofC. kefyr 23 and L. lactis subsp.lactisbiovar. diacetylactis C1 could be too high for ac-ceptable fermented milk. It would be of interest tostudy the effect of such high acetaldehyde levels onthe sensory acceptance of the yeast–LAB fermentedmilk. Acetaldehyde is one of the metabolites pro-duced during alcoholic fermentation by yeastsŽ .Walker, 1998 . Acetaldehyde is also formed bysome lactobacilli from threonine and during

Ž .metabolic breakdown of sugars Dellaglio, 1988 . Ifthe yeast excretes threonine into the milk, this mayresult in increased levels of acetaldehyde in co-cul-

Ž .ture Rysstad et al., 1990 .L. lactis subsp. lactisbiovar. diacetylactis C1 was the only strain thatproduced diacetyl in both single and co-culture. Thesediacetyl levels were reduced in the presence ofC.tropicalis 78 indicating possible production of theacetoinrdiacetyl reductase enzyme by this yeaststrain.

Ethanol and carbon dioxide were very importantin determining properties of milk fermented withC.kefyr. Carbon dioxide production in yeasts is linkedto the glycolytic breakdown of carbohydratesŽ .Walker, 1998 . Apart from the sensory propertiesimparted to fermented milk, high carbon dioxidelevels may lead to bursting of some types of packag-ing such as plastic sachets. In Zimbabwe, fermentedmilk is normally sold in plastic sachets.C. kefyr 23produced high levels of ethanol in both single andco-culture, comparable to those found in kefir, a

Žyeast–LAB fermented product Duitschaever et al.,.1987; Tamime and Marshall, 1997 , and koumiss

Ž .Marshall, 1987; Mann, 1989 . This could suggestsimilarities of our yeast–LAB products with kefirand koumiss.

Although no specific effect of pyroglutamic acidon yeast strains could be directly deduced or inferredfrom this study, the organic acid is an importantmetabolite, whose antimicrobial activity has been

Ž .reported Huttunen et al., 1995 . It would be ofinterest in future studies to test the effect of this

compound on selected yeast strains. Pyroglutamicacid in milk may be released from the N-terminus ofproteins and peptides by the action of pyrrolidine

Ž .carboxyl peptidase Mucchetti et al., 2000 . Otherauthors report that it is produced by cyclisation of

Ž .glutamic acid by LAB Tschager and Jager, 1988 .Formic acid is also an important metabolite of LABfermentation. Lactococci can produce formate from

Žpyruvate through pyruvaterformate lyase Teuber,.1995 . Additionally, fermentation of citrate leads to a

mixture of products including lactate, acetate, for-Žmate, acetoin, diacetyl and 2,3-butanediol Hugen-

.holtz, 1993 .The similar levels of 2-methyl-1-propanol, 2-

methyl-1-butanol, 3-methyl-1-butanol, acetaldehydeand ethanol produced byL. lactis subsp.lactis bio-var. diacetylactis C1 andS. cereÕisiae 71, L. lactissubsp.lactis biovar. diacetylactis C1 andC. collicu-losa 41, L. lactis subsp.lactis biovar. diacetylactisC1 andS. dairenensis 32, andL. lactis subsp.lactisbiovar. diacetylactis C1 and D. bruxellensis 43 co-cultures seem to suggest that the yeast strains in-volved may have similar metabolic patterns in theUHT milk. From taxonomic studies, these yeaststrains could assimilate lactate and galactose, but not

Ž .lactose Gadaga et al., 2000 . These co-cultures pro-duced more of the malty compounds, acetaldehydeand ethanol than the single yeast or LAB cultures,suggesting some form of interaction. Lactic acidbacteria are auxotrophic for some amino acids sothat if the yeasts produce these essential amino acids,

Ž .growth could be enhanced Teuber, 1995 . Leroi andŽ .Pidoux 1993b noted an interaction between yeasts

and LAB, where they suggested that the yeasts couldprovide vitamins, amino acids and growth factors forbacteria, while the bacterial end products could beused by the yeasts as an energy source.

The use of the potential opportunistic pathogenicyeastsC. tropicalis and C. lusitaniae in this studywas justified by the need to establish whether theyhave potential to grow in the UHT milk. It is,however, important that yeast strains selected for useas starter cultures are screened for safety.

5. Conclusion

This study showed that in co-culture, the LABmainly influenced the final pH and metabolite con-


tent of the samples, except for the samples contain-ing C. kefyr 23, which grew in the UHT milk andproduced high levels of flavour compounds includ-ing ethanol and acetaldehyde. The basis for growthof the non-lactose fermenting yeasts in single culturein the UHT milk was less clear, but was potentiallydue to proteolysis and lipolysis. However, furtherinvestigations will be needed to confirm this asser-tion. The higher final populations of some LABstrains in co-culture with yeasts, especiallyC. kefyr23, S. cereÕisiae 71 and S. dairenensis 32, werededuced as an indication of a beneficial effect of theyeasts on the LAB. However, in some cases co-inoc-ulation did not result in a major effect on the finalnumbers of either the yeast or the LAB. Enhancedproduction of flavour compounds such as ethanol,acetaldehyde and malty compounds by some yeast–LAB co-cultures was presumed to be indicative ofinteraction between the yeasts and LAB. These pos-sible interactions could be important in the develop-ment of yeast–LAB fermented milk products.C.kefyr 23 could be a suitable starter culture for milkfermentation.

Acknowledgements

The authors are grateful to Kari R. Olsen forassistance with the headspace-GC and HPLC analy-ses. The Norwegian Universities Committee for Re-

Žsearch, Development and Education NUFU Project.26r96 , through the Agricultural University of Nor-

way and the University of Zimbabwe, funded thisstudy.

References

Anders, R.F., Jago, G.R., 1970. The concentration of oleic acid incheddar cheese relative to loss of starter viability. Aust. J.Dairy Technol. 25, 68–72.

Broome, M.C., Thomas, M.P., Hillier, A.J., Horwood, J.F., Jago,G.R., 1979. The effect of linoleic acid on the growth andmetabolism ofStreptococcus lactis. Aust. J. Dairy Technol.34, 163–168.

Davidson, B.E., Hillier, A.J., 1995. Developing new starters forfermenting milk products. Aust. J. Dairy Technol. 50, 6–9.

Dellaglio, F., 1988. Starters for fermented milks: taxonomy andmetabolism. Bulletin of the International Dairy Fermentation,vol. 227. IDF, Brussels, pp. 7–18.

Duitschaever, C.L., Kemp, N., Emmons, D., 1987. Pure cultureformulation and procedure for the production of kefir. Milch-wissenschaft 42, 80–82.

Fekadu, B., 1994. Present situation and future aspects of milkproduction, milk handling and processing of dairy products inSouthern Ethiopia. PhD thesis, Agricultural University of Nor-

˚way, As, Norway.Feresu, S.B., Muzondo, M.I., 1990. Identification of some lactic

acid bacteria from two Zimbabwean fermented milk products.World J. Microbiol. Biotechnol. 6, 178–186.

Fleet, G.H., 1990. Yeasts in dairy products. J. Appl. Bacteriol. 68,199–211.

Fleet, G.H., 1992. Spoilage yeasts. Crit. Rev. Biotechnol. 12,1–44.

Gadaga, T.H., Mutukumira, A.N., Narvhus, J.A., 2000. Enumera-tion and identification of yeasts isolated from Zimbabwean

Ž .traditional fermented milk. Int. Dairy J. 10 7 , 459–466.Hugenholtz, J., 1993. Citrate metabolism in lactic acid bacteria.

FEMS Microbiol. Rev. 12, 165–178.Huttunen, E., Noro, K., Jang, Z., 1995. Purification and identifica-

tion of antimicrobial substances produced by twoLactobacil-lus casei strains. Int. Dairy J. 5, 503–513.

International Dairy Federation, 1990. International standard 94B,Enumeration of Yeasts and Moulds: Colony Count Techniquesat 258C. IDF, Brussels.

Jakobsen, M., Narvhus, J., 1996. Yeasts and their possible benefi-cial and negative effects on the quality of dairy products. Int.Dairy J. 6, 755–768.

Leroi, F., Pidoux, M., 1993a. Characterisation of interactionsbetweenLactobacillus hilgardii and Saccharomyces florenti-nus isolated from sugary kefir grains. J. Appl. Bacteriol. 74,54–60.

Leroi, F., Pidoux, M., 1993b. Detection of interactions betweenyeasts and lactic acid bacteria isolated from sugary kefirgrains. J. Appl. Bacteriol. 74, 48–53.

Ž .Mann, E.J., 1989. Kefir and koumiss. Dairy Ind. Int. 54 9 , 9–10.Marshall, V.M., 1987. Fermented milks and their future trends: I.

Microbiological aspects. J. Dairy Res. 54, 559–574.Marshall, V.M.E., Tamime, A.Y., 1997. Physiology and biochem-

Ž .istry of fermented milks. In: Law, B.A. Ed. , Microbiologyand Biochemistry of Cheese and Fermented Milk. BlackieAcademic and Professional, London, pp. 153–192.

Montanari, G., Zambonelli, C., Grazia, l., Kamesheva, G.K.,Shigaeva, M.K.H., 1996.Saccharomyces unisporous as theprincipal alcoholic fermentation microorganism of traditionalkoumiss. J. Dairy Res. 63, 327–331.

Mucchetti, G., Locci, F., Gatti, M., Neviani, E., Addeo, F.,Dossena, A., Marchelli, R., 2000. Pyroglutamic acid in cheese:presence, origin and correlation with ripening time of GranaPadano cheese. J. Dairy Sci. 83, 659–665.

Mutukumira, A.N., 1995. Properties of amasi, a natural fermentedmilk produced by smallholder milk producers in Zimbabwe.Milchwissenschaft 50, 201–205.

Mutukumira, A.N., 1996. Investigation of some prospects for thedevelopment of starter cultures for industrial production oftraditional fermented milk in Zimbabwe. PhD Thesis, Agricul-

˚tural University of Norway, As, Norway.


Narvhus, J.A., Thorvaldsen, K., Abrahamsen, R.K., 1990. Quanti-tative determination of volatile compounds produced byLac-tococcus spp., using automatic, direct head-space gas chro-matography. Dairying in a Changing World: XXIII Interna-tional Dairy Congress. IDF, Brussels, 522.

Narvhus, J.A., Hulbækdal, A., Thorvaldsen, K., Abrahamsen,R.K., 1991. Measurement of CO production and O2 2

metabolism by pure and mixed cultures of lactic acid bacteriagrowing in milk. Actes du Colloque Lactic ’91. Centre dePublications de l’Universite de Caen, Caen, France, 371.

Narvhus, J.A., Østeraas, K., Mutukumira, T., Abrahamsen, R.K.,1998. Production of fermented milk using a malty compoundproducing strain ofLactococcus lactis subsp. lactis biovar.diacetylactis isolated from Zimbabwean naturally fermentedmilk. Int. J. Food Microbiol. 41, 73–80.

Peppard, T.L., Halsey, S.A., 1981. Malt flavour-transformation ofcarbonyl compounds by yeast during fermentation. J. Inst.Brew. 87, 3860–3890.

Roostita, R., Fleet, G.H., 1996. Growth of yeasts in milk andassociated changes to milk composition. Int. J. Food Micro-biol. 31, 205–219.

Rysstad, G., Knutsen, W.J., Abrahamsen, R.K., 1990. Effect ofthreonine and glycine on acetaldehyde formation in goats’milk yogurt. J. Dairy Res. 57, 401–411.

Sarais, I., Piussi, D., Aquili, V., Stecchini, M.L., 1996. Thebehaviour of yeast populations in Stracchino cheese packagedunder various conditions. J. Food Prot. 59, 541–544.

Sserunjogi, M.L., 1999. Ugandan indigenous fermented dairy

products, with particular focus on Ghee. PhD thesis, Agricul-˚tural University of Norway, As, Norway.

Stam, H., Hoogland, M., Laane, C., 1998. Food flavour fromŽ .yeast. In: Wood, B.J.B. Ed. , Microbiology of Fermented

Foods. Blackie Academic and Professional, London, pp. 505–542.

Subramanian, P., Shankar, P.A., 1985. Commensalistic interactionbetween Lactobacillus acidophilus and lactose fermentingyeasts in the preparation of acidophilus-yeast milk. Cult. DairyProd. J. 20, 19–26.

Tamime, A.Y., Marshall, V.M.E., 1997. Microbiology and tech-Ž .nology of fermented milks. In: Law, B.A. Ed. , Microbiology

and Biochemistry of Cheese and Fermented Milk. BlackieAcademic and Professional, London, pp. 57–152.

Tamime, A.Y., Robinson, R.K., 1999. Yoghurt: Science andTechnology. Pergamon, Oxford, England.

Teuber, M., 1995. The genusLactococcus. In: Wood, B.J.B.,Ž .Holzapfel, W.H. Eds. , The Genera of Lactic Acid Bacteria.

Blackie Academic and Professional, London, pp. 173–234.Tschager, E., Jager, H., 1988. Pyroglutaminsaure in milch und¨

milchprodukten unter besonderer berucksichtigung des kases.¨ ¨Milchwirtsch. Ber. 95, 79–83.

Venugopal, V., 2000. Traditional preservatives—vegetable oils.Ž .In: Robinson, R.K., Batt, C.A., Patel, P.D. Eds. , Encyclope-

dia of Food Microbiology. Academic Press, London, pp.1743–1749.

Walker, G.M., 1998. Yeast Physiology and Biotechnology. Wiley,Chichester, England.

the growth and interaction of yeasts and lactic acid bacteria isolated from zimbabwean naturally...

Documents