Microbiological and Rheological Studies on Portuguese Kefir Grains*

International Journal of Food Science and Technology 1996, 31, 15–26 15

Microbiological and rheological studies on Portuguese keﬁr grains*

Manuela E. Pintado, J. A. Lopes Da Silva, Paulo B. Fernandes, F. Xavier Malcata§ & Tim A. Hogg

Escola Superior de Biotecnologia, Universidade Católica Portuguesa, Porto, Portugal

Summary The native bacteria and yeasts present in Portuguese kefir grains stored under four distinct sets of environmental conditions have been isolated and identified on the basis of morphology and biochemical tests. The microbial population of the kefir grains as a whole has been characterized in terms of rates of biomass production and formation of lactic acid and ethanol. The rheological properties of the purified polysaccharide (kefiran) produced by the microflora of the grains and accumulated therein were studied in a low water activity solvent and as a component of a binary gel containing either ␬-carrageenan or xanthan gum.

Keywords Biochemical parameter, dairy foods, gum blend behaviour, keﬁran, mechanical spectra, storage and loss moduli.

(previously included in L. brevis) (La Reviere et Introduction al., 1967; Kandler & Kunath, 1983), although Kefir is a fermented beverage which has its several other varieties can be found, e.g. bacteri- ancient origins in the Caucasian mountains. This al species of such genera as Streptococcus, light alcoholic beverage is prepared by inocula- Leuconostoc and Acinetobacter (Rosi, 1978b; Rosi tion of raw milk with existing kefir grains (which & Rossi, 1978), and mould species such as can be stored in dried form). Kefir grains exhibit Geotrichum candidum (Kosikowski, 1977). At an irregular surface and they consist of a gel least one quarter of the dry matter of the kefir matrix in which yeasts and bacteria are imbedded grains consists of a capsular polysaccharide and live symbiotically. usually denominated kefiran (La Reviere et al., The microflora of kefir grains, which depends 1967). Kefiran is considered to be a result of on their origin (Ottogalli et al., 1973), storage the metabolism of such capsular bacteria conditions and handling pattern (Zourari & as Lactobacillus kefis (La Reviere et al., 1967; Anifantakis, 1980), consists predominantly of Yokoi et al., 1990), a homofermentative, atypi- Saccharomyces delbrueckii and Lactobacillus kefir cal Streptobacterium (Rosi & Rossi, 1978), Lactobacillus kefiranofaciens (Toba et al., 1987; Fujisawa et al., 1988), and a new LAB strain ten- §Correspondent: Escola Superior de Biotecnologia, Rua Dr. António Bernadino de Almeida, 4200 Porto, tatively named KPB-167B (Yokoi et al., 1991). Portugal. Fax: 351 2 590351. Kefiran consists of approximately equal propor- *Part of the data presented in this paper was included in tions of galactose and glucose residues (Kooiman, a poster entitled Studies on kefir grains: microbiological 1968). Rheological studies have shown that characterization, biomass increase rate, and rheology of the kefiran itself has a very low viscosity in solution polysaccharide produced, which was presented at the Prof. Nicolau van Uden Portuguese Symposium of Applied and is unable to form rigid gels in the absence of Microbiology and Biotechnology – First Symposium on ethanol (Mukai et al., 1990, 1991). Although the Yeasts, February 1993, Luso, Portugal. structural characteristics and hydrodynamic prop-

erties of kefiran raise some restrictions on its media, and incubated under different conditions commercial use as a food thickening or gelling depending on the microorganism under study: agent, some useful applications as a food additive (a) for Lactobacillus sp., the medium used was might be devised. However, to date little work MRS (LabM, UK) supplemented with 3% (w/v) has been devoted to the rheological properties of ethanol and 0.5% (w/v) cycloheximide (Sigma, this polysaccharide. USA) and incubation took place anaerobically The purpose of the research reported in this (via Gas Pack BBL™ from Becton Dickinson, communication was to gain insight into the USA) at 30ЊC for 5 days; (b) for Streptococci, the microbiology of Portuguese kefir grains in terms medium used was M17 Agar (Merck, Germany) of identification of their microflora and charac- with 0.5% (w/v) cycloheximide, with aerobic incu- terization of the biomass growth and product bation at 30ЊC for 2 days; (c) for yeasts, the medi- formation rates thereof, and to compare the um used was MEA (Merck) incubated, results thus obtained with results available in aerobically, during 5 days at 25ЊC; (d) for acetic the literature pertaining to kefir grains from acid bacteria, the medium used was WLNA other countries. Several rheological studies were (Oxoid, UK) with 0.5% (w/v) cycloheximide, with also conducted on the polysaccharide extracted aerobic incubation at 25ЊC for 21 days. Selected from the kefir grains in attempts to improve its colonies were purified prior to identification. All potential applications as a gelling agent for isolated strains were stored in the presence of foods. 30% glycerol at Ϫ80ЊC.

Materials and methods Identification of microoganisms Microbial strains For biochemical profiles of lactic acid bacteria The kefir grains (0.7 to 1.4 g fresh, ϳ10% dry basal MRS media without meat extract and glu- matter) were obtained from private sources in cose were used according to Bergey (1986). The Portugal. Propagation and activation was per- inoculated medium was added to the galleries of formed by consecutive immersion (c. 10–12% w/v) API 50™ (Bio Mérieux, France). The additional in fresh batches of sterilized cow’s milk at room tests for streptococci and lactobacilli were exam- temperature three times a week. ined according to the schemes of Harrigan & The pure strains of Lactobacillus brevis NCFB McCance (1976). For lactobacilli the genus was 1749, Lactobacillus kefir NCFB 2753 and further confirmed by SDS-PAGE of the disrupt- Lactococcus lactis subsp. lactis NCFB 604 were ed cells following preliminary growth under stan- obtained from the National Collection of Food dard conditions according to the method of Bacteria (UK). Laemmli (1970) modified by Kiredjian et al. (1986) using chemicals from BioRad (USA); numerical analysis of the protein electrophore- Microbiological analyses gram was performed as described by Kersters & Grains subject to four distinct environmental De Ley (1975). conditions were employed in these experiments, In the case of yeasts, physiological and bio- viz. (i) original grains, (ii) after propagation at chemical characteristics were tested according to home with daily change of milk, (iii) after storage Barnett et al. (1990) and identification was based for 3 months at room temperature (20–25ЊC), on the Yeast Identification PC Program from the and (iv) after refrigerated storage (4ЊC) for 3 same authors. months. For each experiment, 10 g of kefir grains Capsule production (previously washed with sterile water) were homogenized with 90 g of Ringer solution in a For this purpose, two solid media were used, Stomacher LAB-Blender 400™ (Seward Medical, namely KPL, which was prepared as described by UK) for 15 min. The homogenate, after appro- Toba et al. (1986) using lactic acid, glucose, priate decimal dilution, was spread on different galactose, and Tween 80™ (Merck) together with

International Journal of Food Science and Technology 1996, 31, 15–26 © 1996 Blackwell Science Ltd Microbiology and rheology of kefir M. E. Pintado et al. 17 milk whey and white table wine, and a new medi- ethanol following the procedure of Da Silva & um prepared with milk permeate (obtained by Gonçalves (1990). ultrafiltration), added with 10% (w/v) lactose and 3% (v/v) ethanol. Both solid media were supple- Determination of intrinsic viscosities mented with 0.5% (w/v) cycloheximide, inoculated with the homogeneous extracts of kefir grains, Several aqueous solutions of kefiran were and incubated at 30ЊC for 5 days under anaero- prepared so as to give relative viscosities from bic conditions. about to 1.2 to 2.1. Viscosity was measured at Fresh permeate (without agar) was, after 25.0 Ϯ 0.1ЊC with a Cannon-Fenske capillary vis- filter-sterilization (Nucleopore, USA), employed cometer (Hipex, Portugal). The intrinsic viscosity, as the broth medium. After pouring 15 ml of [␩], was obtained by combined application of the broth into screw-capped tubes, each tube was Huggins and Kraemer extrapolations (Morris & inoculated with some of the lactobacilli isolated Ross-Murphy, 1981). from the colonies in MRS added with 3% (v/v) ethanol and incubated at 30ЊC with continuous Preparation of the sucrose/kefiran systems shaking (100 r.p.m.) for 6 days. The kefiran ([␩] ϭ 0.29 m3 kgϪ1) was previously dispersed in water at room temperature under Growth rate moderate shaking for 1 h, heated at 90ЊC for 30 Nine wet grains of activated kefir, with similar min, and then centrifuged for 1 h at 28 000 g. The and representative appearance, were selected and aqueous dispersion of kefiran was heated in a washed with sterile, distilled water. They were paraffin bath to 105ЊC for 3 min, and then dried to constant weight at 25ЊC. Each grain was sucrose (a water activity modifier) was added then submerged in 80 ml of sterile milk in a under continuous stirring. The heating was main- different beaker, and the beakers were incubated tained for another 7 min, water lost being at 22ЊC under gentle shaking conditions. For replaced with fresh deionized water, and the sam- biomass assays the grains were removed, each one ple was transferred to the plate of the instrument at a different time, and dried to constant weight at the desired temperature. at 25ЊC. Assays for DL-lactic acid and ethanol by enzymic methods (Boehringer Manheim, Preparation of the binary solutions of kefiran Germany) of the supernatant milk of the corre- and ␬-carrageenan or xanthan sponding beaker were performed at the six sampling times after the (previously determined) The kefiran and ␬-carrageenan samples were first lag period of 3 days. dispersed in water under moderate agitation for 1 h, at room temperature, and then heated at 90ЊC for 30 min with stirring. The binary systems were Extraction of polysaccharide from the kefir prepared by mixing, at 90ЊC, the solutions of grains kefiran and ␬-carrageenan at the 1:4 (w/w) ratio For the extraction operation, 50 g of grains, for a final 1% (w/w) total concentration of poly- previously washed with sterile water, were kept mer. The hot mixture was maintained at 90ЊC for submerged in 100 ml water at 80ЊC for 1 h 15 min, and then cooled to 55ЊC. In these experi- with uniform stirring (100 r.p.m.). After this ments, the only potassium ions present were those period, the mixture was cooled and centrifuged at accounted for by the salt form of the native ␬-car- 12 000 g for 20 min. The procedure was repeated rageenan. once again with the remaining sediment. The In the case of the kefiran/xanthan mixture, both polysaccharides dissolved in the supernatant were the kefiran and xanthan were first dispersed in a precipitated out by addition of an equal volume 0.1 M solution of sodium chloride under moderate of cold ethanol. The precipitate was redissolved in agitation for 60 min at room temperature, and 100 ml of water and a second purification step then heated to 90ЊC for 30 min with continuous was performed using isopropanol in place of stirring. The binary systems were prepared by

mixing, at 90ЊC, the solutions of kefiran and xan- of 18ЊC hϪ1 (kefiran/␬-carrageenan); in the case of than at the 1:1 (w/w) ratio for a final 1% (w/w) kefiran/xanthan, the mixture was cooled from 70 total polymer concentration, and then maintained to 25ЊC and then reheated to 70ЊC at the rate of at 90ЊC for 30 min under continuous stirring. 60ЊC hϪ1.

Rheological measurements Results and discussion Dynamic rheological measurements were per- Microbiological characterization formed using a Carri-Med CS-50 (UK) con- trolled-stress rheometer with a cone-plate device The proportions of the different groups of (radius 25 mm, cone-plate angle 4Њ) or a parallel microorganisms (bacteria such as lactobacilli and plate geometry (radius 60 mm, gap 4 mm) with streptococci, and yeasts) found in the grains are radial grooves in order to avoid gel slippage. The depicted in Fig. 1. The grains did not reveal the strain amplitude was fixed at 3%. After the hot presence of acetic acid bacteria as reported by sample was transferred to the rheometer plate, the Rosi (1978b). While lactobacilli dominate the exposed surface of the sample was covered with a microflora of the grain, yeasts were apparently thin layer of low viscosity paraffin oil to prevent the least affected by environmental conditions. It evaporation of solvent. is interesting to note the similarities of the micro- Three types of experiments were then per- bial counts of grains 1 and 4, confirming that formed: cure experiments, mechanical spectra, and storage of wet grains at 4ЊC is a good method for cooling/heating cycles. For the cure experiments, the preservation of the grains as generally accept- the kinetics of gel formation at different tempera- ed (Kosikowski, 1977). This result is independent tures was monitored by measuring the storage of the observed increase in weight of the refriger- (GЈ) and the loss (GЉ) moduli at a frequency of 6.3 ated grains, which is consistent with the results rad sϪ1. Mechanical spectra were recorded in a obtained by Toba et al. (1990). constant strain mode over the frequency range After screening bacteria showing different 0.063 to 63 rad sϪ1. For the cooling/heating cycles, colony morphologies on MRS, two colonies the storage modulus, GЈ, and the loss modulus, (denoted hereafter as Isolate I and Isolate II) were GЉ, were obtained from temperature sweep exper- further characterized. Table 1 shows sugar fer- iments at 6.3 rad sϪ1, by cooling the systems from mentation profiles of these two strains, and a 55 to 5ЊC and then reheating to 55ЊC at the rate comparison with those of reference strains. In a

Figure 1 Number of Colony Forming Units (CFU) per unit mass of keﬁr grain for (1) original grains, (2) after propagation at home under typical conditions, (3) after storage for three months at room temperature, and (4) after refrigerated storage (4ЊC) for three months.

Table 1 Results of the tests performed on the microbial isolates

Test Lactobacilli Streptococci Yeasts

Isolate Isolate L. brevis L. keﬁr Isolate Lact. Isolate Sacch. I II III lactis IV unisporus amygdalin ϪϪϪϪϩϩ L-arabinose ϩϩϩϩϪϪϪϪ erythreitol ϪϪϪϪϪϪϪϪ celobiose ϪϪϪϪϩϩϪϪ esculin ϪϪϪϪϩϩ fructose ϩϩϩϩϩϩ galactose ϩϩϩϩϩϩϩϩ glucosamine ϪϪ glucose ϩϩϩϩϩϩϩϩ gluconate ϩϩϩϩϩϪϪϪ glucoronate ϪϪ glycerol ϪϪϪϪϪϪϪϪ inositol ϪϪϪϪϪϪϪ 2-keto-D-gluconate ϪϪϪϪϪϪϪϪ lactate ϪϪ lactose ϩϩϩϩϩϩϪϪ maltose ϩϩϩϩϩϩϪϪ mannitol ϪϪϪϪϪϪϪϪ mannose ϪϪϪϪϩϩ melezitose ϪϪϪϪϪϪϪϪ melibiose ϩϩϩϩϪϪϪϪ Me-␣-D-glucosamine ϪϪ rafﬁnose ϪϩϪϪϪϪϪϪ rhamnose ϪϪϪϪϪϪϪϪ ribose ϩϩϩϩϩϩϪϪ sorbose ϪϪ salicin ϪϪϪϪϩϩ sorbitol ϪϪϪϪϪϪ sucrose ϪϪϩϪϩϩϪϪ trehalose ϪϪϩϪϩϩϪϪ xylose ϪϪϩϪϩϩϪϪ

NH3 from Arg ϩϩϩϩ heterofermentative ϩϩϩϩ growth @ 10ЊC ϩϩ growth @ 40ЊC ϩϩ growth @ 45ЊC ϩϩϩϩϪϪ growth @ 4% NaCl ϩϩ growth @ 6.5% NaCl ϪϪ growth @ pH 9.2 ϩϩ growth @ pH 9.6 ϪϪ 0.01% cycloheximide ϩϩ 0.1% cycloheximide ϩϩ acetic acid production ϪϪ L-lysine ϩϩ

similar fashion, one colony from M17 (denoted strains of lactobacilli detected are heterofermenta- hereafter as Isolate III) and one colony from MEA tive and produce ammonia from arginine. (denoted hereafter as Isolate IV) were further Inspection of Table 1 also indicates that the two characterized, and the results thereof are also strains of Lactobacillus tested diﬀer in only one included in Table 1. As shown in this table, all type of sugar fermented, raﬃnose (which has a

positive result for Isolate II). When compared with Capsule production reference strains of lactobacilli commonly found in All lactobacilli isolated from kefir grains proved kefir grains, it can be observed that L. kefir is the to be encapsulated. Although capsule production most similar to the isolated strains. Confirmation in solid media was not successful, the isolates of the above results was achieved via generation of identified as L. kefir incubated in milk permeate electrophoretic patterns of the whole protein broth for 6 days produced capsules surrounding inventory of the two different isolates and refer- their rod-shaped body which could be easily visu- ence strains. The associated dendograms have alized via Indian ink staining. Further character- shown 94.1% and 89.0% homology with L. kefir ization of the capsular polysaccharide was not for Isolates I and II, respectively; these results are pursued (because it would have required sophisti- in agreement with the conclusions of Kandler & cated analytical techniques that were not avail- Kunath (1983). Different results were obtained by able at our premises at that time); however, our other authors, e.g. L. brevis (now included in L. result is expected in view of the conclusions by La kefir) was isolated by Rosi & Rossi (1978), where- Reviere et al. (1967), viz. that L. kefir (formerly as Ottogalli et al. (1973) described L. brevis and L. L. brevis) was most probably responsible for pro- acidophilus for grains from Bulgaria, Yugoslavia ducing kefiran, the material of the capsules. and Russia, and Molska et al. (1983) isolated L. casei from Polish and Norwegian grains. With respect to lactococci, Isolate III was Growth rate identified as Lactococcus lactis on the basis of the It took almost 3 days before the dry grains sub- sugar fermentation characteristics and other merged in milk started their exponential growth growth conditions (Bergey, 1986). From the great (see Fig. 2). (This lag phase was expected in view degree of homology with the corresponding refer- of the commonly accepted idea that microorgan- ence strain as apparent in Table 1, the isolated isms in dried form take a considerable time to strains could be identified as Lactococcus lactis readapt their metabolic system to biomass pro- subsp. lactis. This result is similar to that duction.) The exponential curve corresponded to obtained by Ottogalli et al. (1973) (i.e. Str. lactis a specific growth rate of ␮ϭ0.002475 hϪ1 (dou- subsp. lactis), but different from that by Rosi & bling time of 16.8 days). Although our result is Rossi (1978) who isolated Str. durans. not of the same order of magnitude of that of The isolated yeasts (Isolate IV) were identified Rozhkova (1984), viz. doubling time of 24 h, it as Saccharomyces unisporus or, synonymously, should be emphasized that the two sets of data Saccharomyces delbrueckii, which is the yeast are difficult to compare in view of the different most frequently isolated from kefir grains (La experimental conditions employed. Reviere et al., 1963). The identification was based The rates of production of lactic acid and on analysis of the results pertaining to the uti- ethanol were followed along with the biomass lization of several sugars (see Table 1) as well as increase, and the results obtained suggest that additional tests of growth in 0.1% (w/v) cyclo- both products are growth-associated (see Fig. 2); heximide, growth on L-lysine, and acetic acid pro- although it is apparent that the concentration of duction, and employed the PC Program written lactic acid and the biomass concentration reach by Barnett et al. (1990) as an identification aid. maxima after the maximum concentration of Although the results obtained for Portuguese ethanol is attained, the time delay in question grains in terms of yeast content compare well cannot be considered to be significant given the with those by Rosi (1978a) and Ohara et al. intrinsic variability of the data presented. (1977), they are quite different from the results available for Russian (i.e. Sacch. lactis), Rheological characterization of the Yugoslavian (i.e. Candida tenuis, Saccharomyces kefiran/sucrose systems lactis, and Sacch. carlbergensis) and Bulgarian grains (i.e. Cand. pseudotropicalis) by Ottogalli et The results of the gel cure experiments performed al. (1973), and those reported by Kosikowski on the kefiran/sucrose systems (1% kefiran, 60% (1977) (i.e. Sacch. kefir and Torula kefir). sucrose) at some ageing temperatures are depicted

Figure 2 Growth data (᭺——᭺) 140 for individual keﬁr grains each in 80 ml of milk at 22ЊC under 130 gentle shaking conditions superimposed on the curves of 120 release of ethanol (ᮀ——ᮀ) and lactic acid (᭝——᭝). 110 100 90

Biomass dry matter (mg) 80 70

in Fig. 3. The general behaviour is typical of a gels changes with the ageing temperature to which biopolymer gel, with both moduli increasing as a the system has been exposed (Fig. 4). The behav- result of the increasing junction zones density, iour is characteristic of a weak biopolymer gel and with the elastic component (storage modulus, with low enough density of junction zones to GЈ) rising rapidly at first and then more slowly as allow some molecular rearrangements to take the apparent plateau is approached. Note the place within the network over the time scale under small difference between the dynamic moduli, consideration. which is a characteristic of weak gels. The time elapsed from the moment when the sample reach- Rheological characterization of the es the temperature under study until the kefiran/␬-carrageenan binary systems cross-over of the storage (GЈ) and loss (GЉ) moduli can be considered as the gel time. The effect of addition of kefiran on the thermal The viscoelastic behaviour of the kefiran/sucrose behaviour of ␬-carrageenan systems is shown in

Figure 3 Storage modulus, GЈ (open symbols), and loss Figure 4 Mechanical spectra for a 1% kefiran gel (60% modulus, GЉ (filled symbols), as a function of time for a (w/w) sucrose) measured at 16 h ageing time, for different 1% kefiran sol/gel (60% (w/w) sucrose) at different ageing ageing temperatures: 10ЊC (᭝ or ᭡), 20ЊC (ᮀ or ᭿), and temperatures: 10ЊC (᭛ or ᭜), 20ЊC (ᮀ or ᭿), and 30ЊC 30ЊC (᭺ or ᭹). Open symbols denote the storage (᭺ or ᭹). (For clarity, only some of the experimental modulus, GЈ, whereas filled symbols denote the loss points obtained are shown.) modulus, GЉ.

Figure 5 Evolution of GЈ (᭺ for cooling and ᭹ for heating) and GЉ (for ᮀ cooling and ᭿ for heating) as functions of temperature for 1:4 keﬁran/ ␬-carrageenan at 1% total polymer concentration for the frequency of 6.3 rad sϪ1. Tg: gelation temperature;

Tm: melting temperature.

Fig. 5 as the evolution of GЈ and GЉ versus tem- 15ЊC, GЈ increased rapidly at first and then more perature for the 1:4 kefiran/␬-carrageenan blend slowly approaching an apparent plateau. The at 1% final polymer concentration. The mixing cross-over of GЈ-GЉ occurred after 20 min. The ratio of 1:4 (kefiran/␬-carrageenan) was used mechanical spectra exhibited by the 1:4 because it allows a direct comparison of the data kefiran/␬-carrageenan blend at 15ЊC are displayed found for the kefiran/␬-carrageenan mixture with in Fig. 6. The fact that GЈϾGЉthroughout the those of 1% 1:4 guar/␬-carrageenan previously frequency range coupled with the fact that both described by Fernandes et al. (1992). A hysteresis GЈ and GЉ are quite dependent upon frequency is clearly exhibited via the gelation temperature confirms that the system abides by the usual

(Tg) of c. 13ЊC and the melting temperature (Tm) deﬁnition of a weak gel (Clark & Ross-Murphy, of c. 21ЊC, assuming that the sol–gel transition 1987). corresponds to the cross-over of GЈ and GЉ The data obtained for 1:4 keﬁran/␬-car- (Cuvelier et al., 1990; Lin et al., 1991). During rageenan blend can be discussed by reference to

maturation of the keﬁran/␬-carrageenan system at the values of Tg, Tm and GЈ at 5ЊC encompassing

Figure 6 Frequency dependence of the viscoelastic moduli, GЈ (᭺) and GЉ (ᮀ), for 1:4 keﬁran/ ␬-carrageenan at 1% total polymer concentration at the temperature of 15ЊC.

Table 2 Parameters related to the sol–gel transition of the concentration in the ␬-carrageenan-enriched binary system constituted by keﬁran and ␬-carrageenan phase. This interpretation is consistent with the (␬-car) results described for the 1:4 guar gum at 1% (see

System Tg (؇C) Tm (؇C) G؅ at 5؇C (Pa) Table 2) (Fernandes et al., 1992, 1993). The difference between GЈ values displayed by the two 0.8% ␬-car 0516 06 mixed systems (kefiran/␬-carrageenan and guar 1% ␬-car 24 32 54 gum/␬-carrageenan) at 5ЊC should be ascribed to 1% 1:4 kefiran/␬-car 13 21 09 1% 1:4 guar/␬-car 10 25 20 the lower intrinsic viscosity of the kefiran sample ([␩] ϭ 0.29 m3 kgϪ1) in contrast to that of guar T : gelation temperature. water g gum (1.2 m3 kgϪ1). Tm: melting temperature. .G؅: storage modulus at the frequency of 6.3 rad nϪ1 Rheological characterization of the kefiran/xanthan binary systems ␬-carrageenan alone at the concentrations of 0.8% and 1% (Table 2). The first result expected The evolution of GЈ and GЉ during the cooling from the addition of kefiran to ␬-carrageenan and heating cycles of the 0.5% and 1% xanthan 3 Ϫ1 would be a dilution effect since there is a lower- gum ([␩]0.1 M NaCl ϭ 3.3 m kg ), 1% 1:1 kefiran/xan- ing of the ␬-carrageenan concentration down to than gum and 1% 1:1 guar gum/xanthan gum sys- 0.8% in the 1:4 mixed system at 1%. In fact, as tems are depicted in Figs. 7 and 8. The GЈ apparent from inspection of Table 2, Tg, Tm and modulus for 0.5% and 1% xanthan gum increases GЈ for the mixture correspond to an intermediate strongly as the temperature is decreased from 70 behaviour between a ␬-carrageenan concentration to 25ЊC. The GЈ vs. temperature profiles for cool- of 0.8% and 1%. This means that kefiran does not ing and heating were virtually the same. The GЉ exhibit synergistic effects when mixed with ␬-car- vs. temperature profile was essentially flat in this rageenan similar to those described for mixtures temperature range, and again the heating and of locust bean gum/␬-carrageenan (Fernandes et cooling cycles gave identical results. In the pres- al., 1991); however, kefiran plays a role in the ence of salt (0.1 M NaCl) the xanthan molecules gelation of the system. This behaviour should be should be present in the system in an ordered explained by a stabilizing effect of kefiran in the state (Milas & Rinaudo, 1986); hence, the GЈ and system which is originated by an increase in the GЉ vs. temperature profiles of the ordered xan- actual concentration of ␬-carrageenan due to vol- than molecules can be associated with the degree ume exclusion, which results in a higher actual of macromolecular association which increases

Figure 7 Evolution of GЈ as a function of temperature at the frequency of 6.3 rad sϪ1 for xanthan at the concentrations of 0.5% (ᮀ) and 1% (᭺), and binary mixtures of xanthan with guar (᭡) and with keﬁran (᭜).

Figure 8 Evolution of GЉ as a function of temperature at the frequency of 6.3 rad sϪ1 for xanthan at the concentrations of 0.5% (ᮀ) and 1% (᭺), and binary mixtures of xanthan with guar (᭡) and with keﬁran (᭜).

with decreasing temperature. As in the case of 1% total polymer concentration. Kefiran/xanthan xanthan alone, both 1% 1:1 kefiran/xanthan or mixtures displayed similar trends to those of the guar/xanthan mixtures displayed an increase of guar/xanthan blend; however, the GЈ and GЉ vs. GЈ as the temperature dropped from 70 to 25ЊC. temperature profiles are lower than in the case of The GЉ vs. temperature profile was flat for the 1% the guar/xanthan blend. By the same token, the 1:1 kefiran/xanthan mixture, whereas in the case lower GЈ and GЉ values described for the of 1% 1:1 guar/xanthan blend a slight increase of kefiran/xanthan mixture when compared with GЉ was observed with the decrease of tempera- that of guar/xanthan can also be explained by the ture. (The GЈ vs. temperature profiles of both lower hydrodynamic volume occupied by the blends for either cooling or heating were again kefiran molecules in solution compared with the essentially the same.) In general, the GЈ and GЉ guar molecules (a fact that can be accounted for vs. temperature profiles of both kefiran or by the different values of the intrinsic viscosity of guar/xanthan mixtures displayed a thermorheo- guar and kefiran samples). logical behaviour which is intermediate between those of 0.5% and 1% xanthan cooling-heating Conclusions cycles. In the present experimental conditions (0.1 M NaCl), no significant interactions should take The results reported in this communication per- place between guar and xanthan (Lopes et al., taining to Portuguese native kefir grains are in 1992). Therefore, the difference between the GЈ general agreement with those obtained for kefir and GЉ vs. temperature profiles observed for grains from other countries in respect of the kefiran or guar/xanthan mixtures and those for microbial flora, especially lactobacilli and lacto- 0.5% and 1% xanthan should be ascribed to a cocci. lack of interactions between kefiran (or guar) and Kefiran can form weak gels in conditions of xanthan. Consequently, the differences described low water activity with potential practical appli- between the GЈ and GЉ vs. temperature profiles of cations. The effect of kefiran on the thermal the two mixtures (kefiran or guar/xanthan) and behaviour of ␬-carrageenan or xanthan systems those of xanthan alone should reflect the lower shows no synergistic effects. The thermorheologi- intrinsic viscosities of kefiran and guar when com- cal profiles of kefiran/xanthan or ␬-carrageenan pared with that of the xanthan gum sample. This mixtures are comparable to those described for fact can also be explained by a dilution effect guar/xanthan or ␬-carrageenan mixed systems. since there is a lowering of the xanthan concen- The studies on the viscoelastic characteristics of tration down to 0.5% in the 1:1 mixed systems at biopolymers as a function of temperature allows

International Journal of Food Science and Technology 1996, 31, 15–26 © 1996 Blackwell Science Ltd Microbiology and rheology of kefir M. E. Pintado et al. 25 an assessment of their performances as ingredi- Kandler, O. & Kunath, P. (1983). Lactobacillus kefir sp. ents in food systems which undergo thermal treat- nov., a component of the microflora of kefir. ment. Systematic Applied Microbiology, 4, 286–294. Kersters, K. & De Ley, J. (1975). Identification and grouping of bacteria by numerical analysis of their electrophoretic protein patterns. Journal of General Acknowledgments Microbiology, 87, 333–342. The experimental contribution of J. António Kiredjian, M., Holmes B., Kersters K., Guilvout I. & De Ley J. (1986). Alcaligenes piechaudii, a new species of Couto (c/o Escola Superior de Biotecnologia) to human clinical specimens and the environment. the performance of the electrophoresis experi- International Journal of Systematic Bacteriology, 36, ments is gratefully acknowledged. The ␬-car- 282–288. rageenan and the guar gum were gifts from SBI Kooiman, P. (1968). The chemical structure of kefiran, (France). The xanthan gum was kindly supplied the water-soluble polysaccharide of the kefir grain. by Jungbunzlauer Xanthan (Austria). White table Carbohydrate Research, 7, 200–211. Kosikowski, F.V. (1977). Cheese and fermented milk foods. wine was a gift by SOGRAPE (Portugal). Milk Pp. 40–42. Ann Arbor, Michigan: Edwards Brothers. was a gift by AGROS (Portugal). Laemmli, U.K. (1970). Cleaving of the structural proteins during the assembly of the head of bacteriophage T4 . Nature (London), 227, 680–685. References La Reviere, J.W.M. (1963). Studies on kefir grain. Journal of General Microbiology, 31, 5. (Proceedings of Bergey (1986). Manual of Determinative Bacteriology. Pp. the Society for General Microbiology). 1043–1065; 1209–1234. Baltimore: Williams & Wilkins. La Reviere, J.W.M, Kooiman, P. & Schmidt, K. (1967). Barnett, J.A., Payne, R.W. & Yarrow, D. (1990). Yeasts: Kefiran, a novel polysaccharide produced in the kefir Characteristics and Identification. Pp. 27–31, 602. grain by Lactobacillus brevis. Archives of Microbiology, London: Cambridge University Press. 59, 269–278. Clark, A.H. & Ross-Murphy, S.B. (1987). Structural and Lin, Y.G., Mallin, D.T., Chien, J.C.W. & Winter, H.H. mechanical properties of biopolymer gels. Advances in (1991). Dynamic mechanical measurement of Polymer Science, 83, 55–192. crystallization-induced gelation in thermoplastic Cuvelier, G., Peigny-Nourry, C. & Launay, B. (1990). elastomeric poly propylene. Macromolecules, 24, Viscoelastic properties of physical gels: critical 850–854. behaviour at the gel point. In: Gums and Stabilizers for Lopes, L., Andrade, C.T., Milas, M. & Rinaudo, M. the Food Industry – 5, (edited by G.O. Phillips, D.J. Wedlock, and P.A. Williams. Pp. 549–552. Oxford: (1992). Role of conformation and acetylation of IRL Press. xanthan-guar interaction. Carbohydrate Polymers, 17, Da Silva, J.A.L. & Gonçalves, M.P. (1990). Studies on a 121–126. purification method for locust bean gum by Milas, M. & Rinaudo, M. (1986). Properties of xanthan precipitation with isopropanol. Food Hydrocolloids, 4, gum in aqueous solutions: role of the conformational 277–287. transition. Carbohydrate Research, 158, 191–204. Fernandes, P.B., Gonçalves, M.P. & Doublier, J.L. Molska, I., Moniuszko, I., Komorowska, M. & (1991). A rheological characterization of Merilainen, V. (1983). Characteristics of bacilli of kappa-carrageenan/galactomannan mixed gels: a Lactobacillus casei species appearing in kefir grains. comparison of locust bean gum samples. Carbohydrate Acta Alimentaria Polonica, 9, 79–89. Polymers, 16, 253–274. Morris, E.R. & Ross-Murphy, S.B. (1981). Chain Fernandes, P.B., Gonçalves, M.P. & Doublier, J.L. flexibility of polysaccharides and glycoproteins from (1992). Effect of galactomannan addition on the viscosity measurements. In: Techniques in Carbohydrate thermal behaviour of kappa-carrageenan gels. Metabolism B310, (edited by D.H. Northcote). Pp. Carbohydrate Polymers, 19, 261–269. 1–46. Elsevier/North-Holland Scientific Publishers, Fernandes, P.B., Gonçalves, M.P. & Doublier, J.L. Amsterdam. (1996). Influence of locust bean gum on the rheological Mukai, T., Watanabe, N., Toba, T., Itoh, T. & Adachi, properties of kappa-carrageenan systems in the vicinity S., (1991). Gel-forming characteristics and rheological of the gel point. Carbohydrate Polymers (in press). properties of kefiran. Journal of Food Science, 56, Fujisawa, T., Adachi, S., Toba, T., Arihara, K. & 1017–1018, 1026. Mitsuoka, T. (1988). Lactobacillus kefiranofaciens sp. Mukai, T., Toba, T., Itoh, T., Nimura, T. & Adachi, S. nov. isolated from kefir grains. International Journal of (1990). Carboxymethyl kefiran: preparation and Systematic Bacteriology, 38, 12–14. viscometric properties. Journal of Food Science, 55, Harrigan, W.F. & McCance, W.F. (1976). Laboratory 1483–1484. Methods in Food and Dairy Microbiology. Pp. 270–277. Ohara N., Kozaki, M. & Kitahara, K. (1977). London: Academic Press. Microbiological studies on kefir I. Isolation and

identification of yeasts from kefir. Journal of KPL medium for polysaccharide prodution by Agricultural Science, 21, 92–103. lactobacillus sp. isolated from kefir grain. Japanese Ottogalli, G., Galli, A., Resmini, P. & Volonterio, G. Journal of Zootechnical Science, 58, 987–990. (1973). Composizione microbiologica, chimica ed Toba, T., Arihara, K. & Adachi, S. (1990). Distribution ultrastruttura dei granuli di kefir. Annuario of microorganisms with particular reference to Microbiologia, 23, 109–121. encapsulated bacteria in kefir grains. International Rosi, J. (1978a). I microorganismi del kefir: Gli lieviti. Journal of Food Microbiology, 10, 219–224. Scienza Tecnica Lattiera-Casearia, 29, 59–67. Yokoi, H., Watanabe, T., & Fujii, Y. (1990). Isolation Rosi, J. (1978b). I microorganismi del kefir: Gli and characterization of polysaccharide-producing acetobatteri. Scienza Tecnica Lattiera-Casearia, 29, bacteria from kefir grains. Journal of Dairy Science, 73, 221–227. 1684–1689. Rosi, J. & Rossi, J. (1978). I microorganismi del kefir: I Yokoi, H., Watanabe, T., Fujii, Y., Mukai, T., Toba, T. fermenti lacticci. Scienza Tecnica Lattiera-Casearia, 29, & Adachi, S. (1991). Some taxonomical characteristics 291–305. of encapsulated Lactobacillus sp. KPB-167B isolated Rozhkova, I.V. (1984). Biomass increase of kefir grains in from kefir grains and characterization of its whey. Molochnaya-Promyshlennost, 9, 14–16. extracellular polysaccharide. International Journal of Toba, T., Abe, S., Arihara, K. & Adachi, S. (1986). A Food Microbiology, 13, 257–264. medium for the isolation of capsular bacteria from Zourari, A. & Anifantakis, E.M. (1980). Le kefir: kefir grains. Agricultural and Biological Chemistry, 50, caractéres physico-chimiques, microbiologiques et 2673–2674. nutritionels. Technologie de prodution. Une revue. Le Toba, T., Abe, S. & Adachi, S. (1987). Modification of Lait, 68, 373–392.

Received 2 August 1993, revised and accepted 15 December 1995