Induction of flocculation in Pediococcus damnosusby different yeast strains

IntroductionMixed microbial cultures play an important role innature and in many industrial processes, especially infood fermentations and in waste water treatment(Harrison, 1978; Verachtert, 1992). These mixed micro-bial populations can be very complex and may consistof bacteria, yeasts and filamentous fungi.

Whereas bacteria/bacteria interactions are well docu-mented (for reviews, see Costerton et al., 1987; Busscheret al., 1993), other interactions such as the yeast/bacteriainteractions or the yeast/fungi interactions are lessstudied and are only poorly understood. Some attentionhas been paid to the yeast/bacterium interaction duringfood fermentations such as beer, kefir and sour-doughfermentation (Kennes et al., 1991; Martens et al., 1991;Leroi and Pidoux, 1993 a,b; Gobetti et al., 1994; Leroiand Courcoux, 1996; Röcken and Vosey, 1995). Thesestudies focus mainly on the microbiological and physi-ological aspects of the mixed cultures.

Little data have been published on the mechanism ofinteraction between yeast and bacteria and its effect onthe stability of the co-culture. White and Kidney (1981)have demonstrated that in the co-sedimentation ofHafnia protea with Saccharomyces cerevisiae during beerfermentation yeast cell wall proteins play an essentialrole. In the case of the Escherichia coli – S. cerevisiae inter-action, evidence has been provided that the bacteria bindto the yeast cell wall by a mannose-specific lectin-typebinding (Ofek et al., 1977; Eshdat et al., 1981; Fironet al., 1982). We have shown that a similar lectin-typeof interaction is responsible for the binding of Pediococcusdamnosus to S. cerevisiae. Moreover, the yeast cell wallprotein that is thought to be involved in the bindinghas been isolated; this protein can induce flocculationof Ped. damnosus cells (Lievens et al., 1994).

In this study, we isolated the flocculation inducingprotein factor from five different yeast strains. A floc-culation assay was developed to quantify the affinity ofthe different flocculation inducing proteins towards thePed. damnosus cells

Material and methodsMicrobial strainsSacchaccharomyces cerevisiae S1 was isolated from a Belgianacidic type of beer. S. cerevisiae 153 and WB are lagerbrewing strains and S. cerevisiae HG is an ale yeast strainform our collection. S. cerevisiae Mn21 is a flocculenthaploid laboratory yeast strain. Pediococcus damnosus12A7 was isolated from Gueuze beer.

Preparation of the Pediococcusstock suspensionPed. damnosus 12A7 was grown in MRS broth (Difco) (deMan et al.,1960) at a temperature of 28°C on a reciprocalshaker at 150 strokes min–1. After 5 days, the pH of thecell culture was adjusted to 7 with 2 M NaOH; the cellswere harvested by centrifugation at 10,000 g for 10 min-utes and resuspended in water at 1/10 of the initialvolume of the culture. This is the Pediococcus stock sus-pension. The stock suspension can be stored during oneweek at 4°C. Due to culture dependent changes in the cellwall properties of the bacteria, absolute figures of floccu-lation can only be compared if the same Pediococcus stocksolution is used, all the measurements described in thisarticle were carried out using the same stock solution.

Production of the yeast culture supernatantYeast was cultured in a 6.2 l fermenter in a minimalmedium containing 0.7% Yeast Nitrogen Base (Difco)and 10 % (w/v) glucose. The fermentations were carriedout at 28°C, pH 4.5 and an agitation speed of 250 revmin–1. The medium was air saturated at pitching, but

11111234567891011112345678920111123456789301111234567894011112345678950111123111

Biotechnology Techniques, Vol 11, No 12, December 1997, pp. 879–884

Biotechnology Techniques · Vol 11 · No 12 · 1997 879

Induction of flocculation in Pediococcusdamnosus by different yeast strainsK. Van den Bremt, D. Iserentant* and H. VerachtertLaboratory for Industrial Microbiology and Biochemistry, K.U. Leuven, Kardinaal Mercierlaan 92, 3001 Heverlee,Belgium. Fax: +32 16 321997, E-mail: [email protected]

The induction of flocculation of Pediococcus damnosus by a proteineous factor, purified from the yeast culture super-natant, was strongly dependent on the yeast strain used. The measurement of flocculation inducing activity can beused to examine the affinity of yeast strains toward eventual bacterial contaminants or to evaluate the stability ofcocultures in mixed culture fermentations.

© 1997 Chapman & Hall

J BT,879-884,930 Bremt 4/12/97 2:03 pm Page 879

no further aeration was applied. The fermentationstopped when the pH increased and the stationarygrowth phase of the yeast cells was reached. The yeastculture was centrifuged at 10,000 g. The pH of thesupernatant was adjusted to pH 7 with 2 M NaOH. Tolimit microbial contamination, 95 % (v/v) ethanol wasadded to reach an ethanol concentration in the culturesupernatant of approximately 10 % (v/v). The culturesupernatant was stored at –20°C until further use.

Polyacrylamide gelelectrophoresisPolyacrylamide gel electrophoresis was carried out on aprecast Tris/glycine 4–20% polyacrylamide gradient gel(Bio-Rad) in a LKB 2050 midget electrophoresis unit.The running buffer contained per litre 3g Tris, 14.4gglycine and 1g SDS.

Measurement of the protein content in the culture supernatant and in the concentrated factor preparationsThe protein content of the samples was measured withthe Bio-Rad-Protein-Assay (Bio-Rad). The standardcurve used to determine the protein content of thesamples from the culture supernatant contained 10 %ethanol, similar to the culture supernatant.

Preparation of the concentrated factorOne volume of culture supernatant was precipitatedwith 3 volumes of methanol and left overnight at 4°C.After centrifugation at 10,000 g, the pellet was dis-solved in a volume of water at 1/10 to 1/20 of the initialvolume, depending on the amount of flocculation factorproduced during fermentation, as measured with theBio-Rad-Protein-Assay.

Measurement of the hexose content of theconcentrated factorThe measurement of the hexose equivalent of theconcentrated factor was based on the colorimetric deter-mination of carbohydrates according to Dubois (1956).Briefly, 200 ml sample was mixed with 200 ml 5 %(w/v) phenol. 1 ml of concentrated (18 M) H2SO4 wasadded and the mixture was vortexed vigorously. Thesolution was then incubated at room temperature during10 minutes and vortexed again. Afterwards, the solu-tion was incubated for another 30 minutes. Theabsorbance was measured at 490 nm. The amount ofhexose is expressed as glucose equivalents.

Flocculation assayThe flocculation assay was essentially carried out asdescribed by Lievens et al. (1994), with minor modifi-cations to improve the reproducibility.

Induction of flocculationFor each yeast strain the flocculation inducing activitywas measured for different concentrations of yeast factor.The flocculation experiments were set up in bottles of24 ml. In each bottle 0.1 to 2 ml of the crude factor wasadded to 200 ml Pediococcus stock suspension. The totalvolume in the bottles was brought to 4 ml by addingwater. For the blank, only water was used to dilute thePediococcus cells. The bottles were shaken at 150 strokesmin–1 in a reciprocal shaker at 28°C for 6 hours.

Measurement of the flocculationAfter incubation the content of the bottles was pouredinto 10 ml glass test tubes and the flocs were allowedto settle for exactly 10 minutes. Then the upper 3.5 mlof the 4 ml suspension was taken and poured in a secondglass tube. Both suspensions were diluted with waterto a total volume of 4 ml. The original glass tubescontaining the settled flocs were vortexed vigorouslyduring 30 seconds to break the flocs and to obtain ahomogenous suspension. The optical density of thesuspension was measured at 600 nm; this value wascorrelated with the biomass in the suspension and thuswith the amount of pediococci that flocculated by theaddition of the yeast factor of the examined strain. TheO.D. of the suspension in the glass tube containing theupper 3.5 ml from the flocculation experiment was alsomeasured spectrophotometrically at 600 nm and corre-lated with the amount of pediococci that did not floc-culate with the yeast factor.

Calculation of the flocculation inducing activityThe flocculation inducing activity is the percentage ofthe pediococci that flocculated on the total amount ofpediococci. This can be described as follows:

Flocculation % = O.D. Flocculated periococciTotal O.D.

with the Total O.D. = O.D. Flocculated pediococci +O.D. pediococci in suspension

Taking the blank sample into account, the absolute floc-culation percentage is:

Flocculation % =Flocculation %sample – Flocculation %blank.

ResultsFlocculation in function of the factor concentrationTwo independent fermentations were carried out usingS. cerevisiae strain S1 (Fermentation A and B). Underthe conditions used, only one major protein band of

11111234567891011112345678920111123456789301111234567894011112345678950111123111

K. Van den Bremt et al.

880 Biotechnology Techniques · Vol 11 · No 12 · 1997

J BT,879-884,930 Bremt 4/12/97 2:03 pm Page 880

about 50.000 daltons can be detected in the culturesupernatant (Fig. 1). After purification of the factor withultrafiltration, evidence was obtained that this proteinwas responsible for the induction of flocculation in Ped.damnosus (Van den Bremt et al., 1997).

The amount of crude flocculation factor, produced in thefermentations was 84.75 and 65.00 mg/ml culture super-natant, respectively. From both culture supernatants, the

yeast factor was precipitated with methanol. For each ofthe concentrated factor preparations, the flocculationinducing activity was measured at different factor con-centrations. The results are presented in Fig. 2. As canbe seen, the flocculation inducing activity increases withincreasing factor concentrations, to reach a saturationplateau at higher factor concentrations. Moreover, theflocculation inducing activity is independent of the fac-tor preparation: the curves obtained in both experimentsshow an almost identical behaviour.

Mathematical expression of the affinityTo obtain a mathematical expression allowing an objec-tive comparison of flocculation factors, produced bydifferent yeast strains, we adapted the Michaelis-Mentenequation. The flocculation inducing activity, obtainedin the test can be considered as a velocity, i.e. theamount of flocs formed during a certain incubation time.The reaction between the yeast factor molecules and thePed. damnosus cells can be described as:

[P] + [F]k+1

←→k–1

[PF]k2

→ [V]

with [P] the concentration of Pediococci; [F] the concen-tration of factor molecules; [PF] the concentration of ahypothetical intermediate and [V] the concentration ofthe sedimented flocs. PF is formed by attachment of fac-tor molecules to the Ped. damnosus cells; this attachmentis reversible. k+1 is the rate constant of the formation ofPF, whereas k-1 the rate constant of the dissociation ofPF into P and F. The formation of PF is dependent uponthe affinity of the factor molecules for the Ped. damnosuscells. Once PF has reached a critical size, the complexwill sediment. We assume that this sedimentation is irre-versible and purely floc size dependent. In that case, therate constant for the sedimentation, k2, will not be influ-enced by the factor – Ped. damnosus affinity and can be considered as a constant in all the experiments.Differences in floc formation velocity will only be due todifferences in the k+1/k-1 ratio.

The floc formation velocity can now be expressed as:

v =d[V]

dt= k2[PF] (1)

The concentration of PF can be calculated from:

d[PF]dt

= k–1[P][F] – k1[PF] – k2[PF]

Assuming a steady state [PF] becomes:

[PF] =k–1[P](k1 + k2)

[F](2)

11111234567891011112345678920111123456789301111234567894011112345678950111123111

Induction of flocculation in Pediococcus damnosus by different yeast strains

Biotechnology Techniques · Vol 11 · No 12 · 1997 881

Figure 2 Flocculation of Ped. damnosus cells, induced bydifferent dilutions of the concentrated flocculation inducingfactor obtained from two fermentations (A: and B:) of S. cerevisiae S1.

Figure 1 SDS-PAGE electrophoresis of the culture super-natant from S. cerevisiae S1 (lane 2), compared with the molecular weight standard (broad range – Bio-Rad) containingmyosin (200 kDa), b-galactosidase (116.25 kDa), phosphory-lase B (97.4 kDa), bovine serum albumine (66.2 kDa), ovalbu-mine (45 kDa), carbonic anhydrase (31 kDa), trypsin inhibitor(21.5 kDa), lysozyme (14.4 kDa) and aprotinin (6.5 kDa).(lane1). The position of the flocculation inducing factor, betweenbovine serum albumine (66.2 kDa) and ovalbumin (45 kDa) isindicated by an arrow.

J BT,879-884,930 Bremt 4/12/97 2:03 pm Page 881

The velocity of the floc formation in (1) is then:

v = k2[PF] =k1k2

k–1 + k2

[P][F] (3)

The maximal speed of the floc formation would be:

vmax = k2[Pt] (4)

with PT the total concentration of Pediococcus cells being:

[Pt] = [P] + [PF] + [V] (5)

Using (2),(4) and (5) in (3) the floc forming speedbecomes:

v =k1[P][F]

(6)vmax [P](k–1 + k2) + k1 [P][F] + [V](k–1 + k2)

which leaves [V ] as the only known concentration inthis equation. Equation (1) is constant when assuminga steady state. Integration of this equation gives:

[V] = [PF]k2t =k1[P]k–1 + k2

[F]k2t (7)

Using (7) in equation (6) the floc forming velocitybecomes:

v = vmax Km +[F]

(1 + k2t)[F](8)

with

Km =k1 + k2

k1

As mentioned above, k2 is considered independent fromthe factor used and constant in all experiments; the vari-ation in Km is purely dependent upon the k-1 / k+1 ratio.

However, although k2 is constant, its value is unknown,which makes that the Km value cannot be experimen-tally deduced using equation 8. Therefore, a new para-meter K′m is defined as the concentration of crude factorat which point half of the flocculation % is reached.Within the experimental set up K′m differs by aconstant factor from Km. The relation is given by

Km = (1 – k2t)K′mThe K′m value can now be used as a degree for theaffinity between the crude factor and the Ped. damnosuscells and can be determined by a classical Lineweaver-Burk plot.

Flocculation inducing activity of different yeast strainsIn order to compare the flocculation inducing activity,fermentations were set up with four different yeast

strains: one haploid lab strain (Mn21), two lagerbrewing strains (S153 and WB) and one ale strain (HG).The results were compared with the results of the alestrain S1. The amount of crude flocculation factor,produced in the different cultures, is summarised inTable 1. For each fermentation, concentrated factor wasprepared by methanol precipitation, and the floccula-tion inducing activity as a function of the factor concen-tration was measured.

Figure 3 shows the flocculation plot for the five differentyeast strains that were examined. For all strains, a satu-ration of flocculation is reached at approximately 70%.However, the factor concentration at which this satura-tion level is reached differs from strain to strain. Basedon these experimental data, Lineweaver-Burk plots wereconstructed, and the K′m value was calculated. Theresults are summarised in Table 2. The factor preparedfrom the two lager strains and from the laboratory strainshows a similar K′m value, whereas the factor from thetwo ale strains clearly has a higher affinity for the Ped.damnosus cells.

DiscussionFor each yeast strain the flocculation inducing activitywas measured as a flocculation percentage as a functionof different concentrations of flocculation inducingfactor. The resulting flocculation plot showed Michaelis-

11111234567891011112345678920111123456789301111234567894011112345678950111123111

K. Van den Bremt et al.

882 Biotechnology Techniques · Vol 11 · No 12 · 1997

Table 1 Concentration of proteins present in thesupernatant of the different yeast strains

Yeast strain Protein concentration in the culture supernatant (m g/ml)

S 1 65S 153 79Mn 21 43HG 144WB 115

Table 2 K′m-values of the flocculation plot from thedifferent S. cerevisiae strains and the correlation of therespective flocculation plot with the Michaelis-Mentenderived Lineweaver-Burk plot by means of R

S. cerevisiae K′m – value Correlationstrain Coefficient R

S1 513 0.981S153 914 0.929Mn21 1085 0.903HG 217 0.980WB 823 0.967

J BT,879-884,930 Bremt 4/12/97 2:03 pm Page 882

Menten saturation kinetics. In all cases, the saturationlevel of the flocculation is around 70%, which indicatethat all the yeast factors bind to the same receptors onthe Ped. damnosus cells and differ only in affinity for thereceptors. For each strain, we could calculate a K′m valueas a measure of the affinity of the respective floccula-tion inducing factor for Ped. damnosus and thus for theflocculation inducing activity of that yeast strain.

The affinity towards the Ped. damnosus cells of the floc-culation factor produced from the lager strains and thelab strain is rather similar. The flocculation factor fromthe ale strains shows a higher affinity for the bacteria. Inthe case of S. cerevisiae S1, this result is expected, becausethis strain was isolated from yeast/bacterium mixed cul-tures. The high affinity of the S. cerevisiae HG factor issomewhat unexpected, although this yeast is consideredas susceptible to contamination in brewing practice.

A comparison of the flocculation inducing activity ofthe different yeast strains should not only take intoaccount the affinity of the factor molecule towards thebacterial cells, but also the amount of factor producedby the yeast during fermentation. HG and S1 factormolecules, for example, show both a very high affinityfor the pediococci but the production of factor duringfermentation is considerably higher for HG than for S1.

Comparison of the data from the five yeast strains showsthat the affinity of the flocculation inducing factor forPed. damnosus cells is strongly strain dependent. This isin agreement with the results obtained by White andKidney (1979), showing that cosedimentation of differ-ent brewing yeasts with beer contaminating bacteria like

Hafnia protea, Lactobacillus, Pediococcus and Acetobacterspecies is strongly dependent on the S. cerevisiae strain.

The method described is this paper may not only beused for measuring the interaction of S. cerevisiae andPed. damnosus but could also be used to examine theaffinity of yeast strains toward other possible contami-nants. A low K′m would be an unwanted feature for ayeast, used in pure culture fermentations, as it can indi-cate a high risk of contamination. For yeasts, used inmixed culture fermentations, however, the low K′m maybe an essential characteristic to stabilize the coculture.

The same method may be useful in the selection of yeaststrains used in the prevention of diarrhea (Surawicz etal., 1989; McFarland et al., 1994). These strains preventthe outgrowth of pathogenic bacteria such as Clostridiumdifficile in the intestinal microflora, probably bycapturing the bacterial binding sites and therebyprecluding the binding of the pathogen to the intestinalcells. The measurement of the K′m may provide anelegant way for the ‘in vitro’ selection of more perfor-mant strains.

ReferencesBusscher, HJ, Geertsma-Doornbusch, GI and van der Mei, HC

(1993). J Appl Bact Symposium Supplement 74:136S-142SCosterton, JW, Cheng, KJ, Geesey, GG, Ladd, TI, Nickel, JC,

Dasgupta, M and Marrie, TJ (1987). Ann Rev Microbiol41:435–464

de Man, JC, Rogosa, M and Sharpe, ME (1960). J Appl Bacteriol23:130–135

Dubois, M, Gilles, KA, Hamilton, JK, Rebers, PA and Smith,F (1956). Anal Chem 28:350–356

Eshdat, Y, Speth, V and Jann, K (1981). Infect Immun 34:980–986

11111234567891011112345678920111123456789301111234567894011112345678950111123111

Induction of flocculation in Pediococcus damnosus by different yeast strains

Biotechnology Techniques · Vol 11 · No 12 · 1997 883

Fig. 3 (a) Flocculation of Ped. damnosus cells, induced by different dilutions of the concentrated flocculation inducingfactor obtained from the fermentations with the S. cerevisiae strains S1 (j), S153 (u), HG (r), Mn21 (s) and WB (●).(b)Lineweaver-Burk plot derived from the experimental data for the flocculation inducing factor isolated from the S. cerevisiaestrains S1 ( ), S153 ( ), HG ( ), Mn21 ( ) and WB ( ).

J BT,879-884,930 Bremt 4/12/97 2:03 pm Page 883

Firon, N, Ofek, I and Sharon, N (1982). Biochem Biophys Res Com105:1426–1432

Gobbetti, M, Corsetti, A and Rossi, J (1994). Appl MicrobiolBiotechnol 41:456–460

Harrison, DEF (1978). Adv Appl Microbiol 24:753–759Kennes, C, Veiga, MC, Dubourguier, HC, Touzel, JP, Albagnac,

G, Naveau, H and Nyns, EJ (1991). Appl Env Microbiol57:1046–1051

Leroi, F and Courcoux, P (1996). J Appl Bacteriol 80:138–146Leroi, F and Pidoux, M (1993a). J Appl Bacteriol 74:48–53Leroi, F and Pidoux, M (1993b). J Appl Bacteriol 74:54–60Lievens, K, Devogel, D, Iserentant, D, Verachtert, H (1994).

Colloids and Surfaces B:Biointerfaces 2:189–198Martens, H, Dawoud, E and Verachtert, H (1991). J Inst Brew

97:435–439McFarland, LV, Surawicz, CM, Greenberg, RN, Fekety, R, Elmer,

GW, Moyer, KA, Melcher, SA, Bowen, KE, Cox, JL, Noorani,Z, Harrington, G, Rubin, M and Greenwald, D (1994). J AmMed Asssoc 271:1913–1918

Ofek, I, Mirelman, D and Sharon, N (1977). Nature 265:623–625Röcken, W and Vosey, PA (1995). J Appl Bacteriol Symposium

Supplement 79:38S-48SSurawicz, CA, Elmer, GW, Speelman, P, McFarland, LV, Chinn,

J and Van Belle, G (1989). Gastroenterology 96:981–988Van den Bremt, K, Nuyens, F, Iserentant, D, Verachtert, H

(1997). Med Fac Landbouww Univ Gent 62/4a: 1185–1192Verachtert, H (1992). Proceeding of the COMETT course on Microbial

Contaminants. Foundation for Biotechnical and IndustrialFermentation Research 7:159–184

White, FH and Kidney, E (1979). Proceedings of the EuropeanBrewing Congress: 801–815

White, FH and Kidney, E (1981). Yeast-bacterium interactionsin the brewing industry. In: Mixed Culture Fermentations, MEBushell and JH Slater, eds pp 121–150. New York: AcademicPress.

11111234567891011112345678920111123456789301111234567894011112345678950111123111

K. Van den Bremt et al.

884 Biotechnology Techniques · Vol 11 · No 12 · 1997

Received 29 August 1997;Revisions requested 5 September 1997;

Final Revisions received 8 October 1997;Accepted 10 October 1997

J BT,879-884,930 Bremt 4/12/97 2:03 pm Page 884