OUP user menu

Pichia anomala and Kluyveromyces wickerhamii killer toxins as new tools against Dekkera/Brettanomyces spoilage yeasts

Francesca Comitini, Jessica Ingeniis De, Laura Pepe, Ilaria Mannazzu, Maurizio Ciani
DOI: http://dx.doi.org/10.1111/j.1574-6968.2004.tb09761.x 235-240 First published online: 1 September 2004


Two yeast killer toxins active on spoilage yeasts belonging to the genus Dekkera/Brettanomyces are here described for the first time. The two toxins produced by Pichia anomala (DBVPG 3003) and Kluyveromyces wickerhamii (DBVPG 6077), and named Pikt and Kwkt, respectively, differ for molecular weight and biochemical properties. Interestingly, the fungicidal effect exerted by Pikt and Kwkt against Dekkera bruxellensis is stable for at least 10 days in wine. Thus, a potential application for the two toxins as antimicrobial agents active on Dekkera/Brettanomyces during wine ageing and storage can be hypothesised.

  • Pichia anomala
  • Kluyveromyces wikerhamii
  • Killer toxin
  • Dekkera/Brettanomyces
  • Spoilage wine yeasts

1 Introduction

Yeasts belonging to the genus Dekkera/Brettanomyces represent a major problem in the wine industry [1]. These yeasts are not normal inhabitants of the grape berry surface and fermenting must, but they can develop in white and red wines at the end of the alcoholic fermentation and during wine ageing in wooden barrels [2,3]. Under these conditions, and dependent on the availability of oxygen [4,5] and carbon and energy sources [6,7], they can produce unpleasant odours and tastes which deeply affect the wine aroma [8,9].

There are various methods that can be used to prevent wine spoilage by Dekkera/Brettanomyces, including: adequate cellar hygiene, sulfitation, and ageing at low temperatures. However, some of these procedures may not be appropriate during wine ageing. In fact, SO2 and low temperatures interfere with the development of both the body and the sensorial properties of wine during ageing, and are therefore detrimental to the quality of the final product. Similarly, wine filtration and barrel steaming, which are generally used for the control of this contamination, may not be recommended; filtration can affect the body and viscosity of aged red wines, and steam treatments are known to damage wooden barrels.

In this context, the exploration of killer yeasts as producers of the mycocins that are able to counteract the activities of these undesired microorganisms in wine, seems interesting. In fact, killer toxins have been already proposed to control spoilage yeasts in the food industry [1012]. Similarly, in the wine industry, killer strains of Saccharomyces cerevisiae are currently used as starter cultures [13,14] and the utilisation of the killer toxin of Kluyveromyces phaffii has been proposed to control apiculate wine yeasts [15].

In this paper, with the aim of developing alternative strategies for the control of Dekkera/Brettanomyces during wine ageing, we have individuated and characterised two novel killer toxins active against these spoilage yeasts.

2 Materials and methods

2.1 Yeast strains and culture media

Yeast strains utilized were obtained from DBVPG Industrial Yeast Collection (Table 1). All the strains were maintained on YEPD (glucose 2%, yeast extract 1%, peptone 2% solidified with agar 1.8% when required) at 4 °C for long term storage and cultivated on YEPD at 25 °C. Killer assay was performed onto malt agar (Difco) buffered with 0.1 M citrate-phosphate buffer (CPB) pH 4.4, unless otherwise stated.

View this table:
Table 1

Yeast strains used

SpeciesStrain designation
Brettanomyces custersianusDBVPG 6709
Brettanomyces naardenensisDBVPG 6712
Brettanomyces spp.DBVPG 4050, 4051, 4052
Candida berthettii KDBVPG 6203 (CBS 5452)
Candida freyschussii KDBVPG 6208 (CBS 2162)
Candida glabrata KDBVPG 6518
Candida membranaefaciens KDBVPG 6145 (CBS 1952)
Candida rugosa KDBVPG 6152 CBS 613
Candida terebra KDBVPG 6028 (CBS 6023)
Candida utilis KDBVPG 6160 (CBS 621)
Debaryomyces hansenii KDBVPG 3505
Debaryomyces polymorphus var. polymorphus KDBVPG 3692
Debaryomyces spp. KDBVPG 3158
Dekkera anomalaDBVPG 6115, 6708, 4075
Dekkera bruxellensisDBVPG 6703, 6704, 6705, 6706, 6707, 6710, 6713
Kluyveromyces dobzhanskii KDBVPG 6074 (UCD 50–45)
Kluyveromyces lactis KDBVPG 6031 (CBS 141)
Kluyveromyces lactis, var lactis KDBVPG 6112 (UCD 70–4)
Kluyveromyces phaffii KDBVPG 6076
Kluyveromyces phaseolosporus KDBVPG 6108 (UCD 50–80)
Kluyveromyces thermotolerans KDBVPG 6272
Kuyveromyces wickerhamii KDBVPG 6077 (UCD 50–210)
Pichia anomala KDBVPG 3650
Pichia anomala KDBVPG 3003
Pichia anomala KDBVPG 3864
Pichia jadinii KDBVPG 6180 (CBS 1600)
Pichia membranaefaciens KDBVPG 3815
Saccharomyces cerevisiae SDBVPG 6500 (NCYC 1006)
Saccharomyces cerevisiae KDBVPG 6567
Saccharomyces cerevisiae KDBVPG 6497 (NCYC 232)
Saccharomyces cerevisiae KDBVPG 6499 (NCYC 738)
Saccharomyces exiguus KDBVG 3192
Williopsis californica KDBVPG 3657
Williopsis saturnus var mrakii KDBVPG 6729 (CBS 1707, NCYC 500)
Williopsis saturnus, var lactis KDBVPG 3127
  • DBVPG: Industrial Yeast Collection, University of Perugia, Italy; CBS: Centraalbureau voor Schimmelcultures, The Netherlands; NCYC: National Collection of Yeast Cultures, UK; UCD: Collection of University of California, Davis, USA. K: Killer strain; S: sensitive strain.

2.2 Killer assays

The preliminary screening was performed by means of streak-plate agar diffusion assay according to Rosini [16]. Approximately 105 cells ml−1 (final concentration) of the strain to be tested for sensitivity to the killer toxin were uniformly suspended in 20 ml Malt Agar (Difco) buffered at pH 4.4 (buffer CPB 0.1 M), maintained at 45 °C in a water bath and immediately poured into sterile Petri dishes. Killer yeast was streaked on the agar surface and plates were incubated at 20 °C for 72 h. Killer activity was evident as a clear zone of inhibition surrounding the streak.

Well test assay was performed according to Woods and Bevan [17]. Briefly, 70 μl of toxin samples were filter-sterilised through 0.45-μm-pore-size membrane filters (Millipore Corp., Bedford, MA) and put into wells (7-mm diameter) cut in the malt agar plates that had previously been seeded with 105 cell ml−1 of a sensitive indicator strain. The killing activity was measured as the diameter of the clear zone of inhibition around the well after incubation for 72 h at 20 °C, and is defined as the mean zone of inhibition of replicate wells.

2.3 Toxin production

A 3-l Biostat B fermenter (Braun) containing 1 l of buffered YEPD was inoculated with 50 ml of a stationary phase starter culture of each strain (DBVPG 3003 or DBVPG 6077), which had been pre-cultured in the same medium. Growth conditions were set as follows: pH, 4.4; temperature, 25 °C; pO2, 20%; stirring at 150 rpm. The supernatants from these DBVPG 3003 and DBVPG 6077 cultures were microfiltered (0.45 μm; Millipore) and concentrated 20-fold by ultrafiltration (stirred ultrafiltration cell, Millipore) through membranes with different cut-offs (10- and 3-kDa-cut-off membranes Amicon, Pharmacia, Uppsala Sweden).

The linear relationship observed between the logarithm of killer toxin concentration and the diameter of the inhibition halo assayed by this well-test method was used to define activity in arbitrary units (AU) [15]. One AU corresponds to the concentration of the toxin contained in 70 μl of the supernatant and able to generate an inhibition halo of 3 mm around the well.

2.4 Protease treatment

Pikt and Kwkt were subjected to protease treatments. Five hundred microlitre of the two toxins (20 AU) was mixed with 125 μl of each of the following proteases: trypsin, papain, proteases XIV and XVIII, and pepsin (10 mg ml−1) and incubated at 25 °C for 24 and/or 72 h. The residual killer activities were evaluated by means of the well-plate assay.

2.5 Evaluation of toxin stability to pH and temperature

pH range of activities of killer toxins was determined as follows: 20 AU ml−1 of each killer toxin were dissolved in 0.1 M CPB (pH 3.0, 3.2, 3.4, 3.6, 3.8, 4.0, 4.2, 4.4, 4.6, 4.8, 5.0). Aliquots of 70 μl were tested in the well-plate assay on malt agar, buffered at the correspondent pHs with 0.1 M CPB. The diameters of the inhibition halos were measured after 2 days of incubation at 25 °C.

Temperature ranges of activities were determined as follows: toxin aliquots were treated at different temperatures (from 20 to 45 °C) for 2 h and used for the well-test assay on Malt Agar buffered with 0.1 M CPB.

2.6 Dose dependent effect of the two toxins

D. bruxellensis DBVPG 6076 (106 cells ml−1) was inoculated into 10 ml of YEPD buffered with 0.1 M CPB that contained increasing concentrations of Pikt or Kwkt. The dose dependence of the two toxins on D. bruxellensis viability was evaluated by means of viable plate counts after 24 h of static incubation at 25 °C.

2.7 Evaluation of toxin activity in wine environment

A starter culture of D. bruxellensis DBVPG 6076 previously adapted in Sangiovese wine, pH 3.76, was inoculated in 10 ml of the same wine to have a final concentration 105 cells ml−1. Different amounts of Pikt and Kwkt (143 and 286 AU) were added to the inoculated wine and the effect of the two toxins on D. bruxellensis DBVPG 6076 was monitored by viable plate count on YEPD. Toxins stability in wine was evaluated by means of well test assay performed after 1, 2, 3, 4, 7, 8, 9, 10 days.

3 Results and discussion

3.1 Screening for yeast strains able to kill Dekkera/Brettanomyces

Twenty-nine killer yeasts were tested for their effects on 15 of the most common Dekkera/Brettanomyces strains found in the wine environment. Twenty-one of the killer strains were not able to kill any of the Dekkera/Brettanomyces target strains. However, as shown in Table 2, the remaining eight killer strains showed activities against B. naardenensis DBVPG 6712, but two of them, P. anomala (DBVPG 3003) and K. wickerhamii (DBVPG 6077), were active against the whole set of Dekkera/Brettanomyces strains screened. Moreover, and as seen previously [18,19], both of these last strains were able to kill the S. cerevisiae DBPVG 6500 (generally used as reference strain for killer toxin sensitivity), thus providing the possibility of studying their modes of action on a known biological model.

View this table:
Table 2

Primary screening of killer yeasts active on Dekkera/Brettanomyces

Dekkera/BrettanomycesKiller strains
Target strains60763650300360286112607760743127
  • 6706 K. phaffii, 3650 P. anomala, 3003 P. anomala, 6028 C. terebra, 6112 K. lactis, var lactis, 6077 K., 6074 K. dobzhanskii, 3127 W. saturnus, var. lactis. + killer activity; − no killer activity.

To further understand the nature of the observed relationship the two strains of P. anomala DBVPG 3003 and K. wickerhamii DBVPG 6077 were cultured in liquid medium and the supernatants were tested for the ability to kill D. bruxellensis DBPVG 6706, type strain of the most common species of the genus in wine [20], and the sensitive strain of S. cerevisiae DBPVG 6500, by means of a well-test assay performed as described previously by Ciani and Fatichenti [15]. It was thus confirmed that the killing action was due to the production of extracellular factors (data not shown). These factors were named Pikt (DBVPG 3003) and Kwkt (DBVPG 6077), from Pitchia anomala and K. wickerhamii killer toxins, respectively.

3.2 Kinetics of growth and killer toxins production

The production of Pikt and Kwkt was monitored during the batch cultivation of the producing strains by means of the well-plate assay. Results showed that DBVPG 3003 starts to produce Pikt after 6 h of incubation, while DBVPG 6077 requires longer for the production of an amount of Kwkt, able to cause a visible inhibition halo around the well starting the production only at the end of the exponential phase (Fig. 1). Well test revealed a constant increase in toxin production during exponential growth with a peak during early stationary phase in both strains. However, while Pikt production was stable during stationary phase, Kwkt concentration diminished with time.

Figure 1

Kinetics of growth and killer toxin production of K. wickerhamii DBVPG 6077 and P. anomala DBVPG 3003. Growth was measured as OD600, toxin production was evaluated as the size of the inhibition halo in well test assays. Error bars indicate ±SD.

3.3 Biochemical characterisation of Pikt and Kwkt

To elucidate the properties of Pikt and Kwkt, particularly in relation to their possible use in wine making, the two toxins were subjected to biochemical characterisation. To optimise their production, P. anomala and K. wikerhamii were grown under controlled conditions in bioreactor and the supernatants were subjected to ultrafiltration in order to concentrate the killer toxins and obtain preliminary information on their molecular weight. It resulted that while Kwkt is retained by a 10-kDa-cut-off membrane, a 3-kDa-cut-off membrane was needed to concentrate Pikt (data not shown), thus demonstrating the differences in the molecular masses of these two toxins.

To confirm the proteinaceous nature of Pikt and Kwkt, they were subjected to protease treatments. As shown in Table 3, both toxins were sensitive to proteolytic enzymes, even though Pikt was completely inactivated only by protease XIV, the most non-specific of the proteases used. Thus the proteinaceous nature of the toxins was confirmed. Subsequently, the effects of both pH and temperature on the activities of Pikt and Kwkt were evaluated. The results are shown in Fig. 2, and they demonstrate that both toxins maintain their killing activities in a pH range compatible with wine making conditions and have a peak of killing activity at pH 4.4. Moreover, Pikt is more resistant to higher temperatures than Kwkt (Fig. 2).

View this table:
Table 3

Pikt and Kwkt sensitivity to protease treatment

Toxin (20 AU)No proteaseTrypsinPapainProtease XIVProtease XVIIIPepsin
Pikt (24 h)++++++++++
Pikt (72 h)+++++++++
Kwkt (24 h)++++
  • ++ standard killer activity; + reduced killer activity; − no killer activity.

Figure 2

Pikt and Kwkt stability at varying pH and temperature. Sensitive strain were (◻) D. bruxellensis DBVPG 6706 (♦) S. cerevisiae DBVPG 6500. Data are given as means ±SD.

The effect of increasing concentrations of Pikt and Kwkt on D. bruxellensis DBVPG 6706 was then tested in buffered YEPD. The results are illustrated in Fig. 3, and show that 28.6 AU ml−1 of Pikt has fungistatic activity, while 57.2 AU ml−1 causes cell death, and hence shows a fungicidal effect (Fig. 3). Kwkt appears to have a stronger killing activity, showing fungicidal effects even at 28.6 AU ml−1.

Figure 3

Dose dependent effect of Pikt and Kwkt on the viability of D. bruxellensis DBVPG 6706. Abscissa insertion indicates viable plate count of DBVPG 6706 immediately after the inoculum (10 × 105 UFC ml−1). Data are representative of two independent experiments.

3.4 Activity of Pikt and Kwkt in wine environment

Co-inoculation of D. bruxellensis and P. anomala or K. wickerhamii in buffered YEPD depresses growth of the target cells. In particular, while a killer/sensitive ratio of 1/10 delays growth of D. bruxellensis cells, a killer/sensitive ratio of 1/1 completely inhibits growth of the sensitive strain (data not shown). Indeed, neither P. anomala or K. wickerhamii have enological aptitude, thus they cannot be used directly during wine ageing and storage. However, both Pikt and Kwkt can be used in wine. Accordingly, the well test assay of the supernatants of Sangiovese wine inoculated with D. bruxellensis and added with Pikt and Kwkt indicates that the two killer toxins maintain the killing action at least for 10 days (Fig. 4). Moreover, viable plate count (Table 4) showed a dramatic decrease in the number of D. bruxellensis cells after 10 days.

Figure 4

PiKt and Kwkt stability in wine. Aliquots of Sangiovese wine inoculated with D. bruxellensis and added with 143 (a) and 286 (b) AU of Pikt or Kwkt were filter sterilised and used in well test assays. The inhibition halos were recorded for up to 10 days. Data are means ±SD.

View this table:
Table 4

Viable plate count of D. bruxellensis DBVPG 6706 in Sangiovese wine after 10 days treatment with Kwkt and Pikt

TrialCFU ml−1
No killer toxin added1.4 × 106
Kwkt (AU 143 ml−1)7.2 × 102
Kwkt (AU 286 ml−1)3.1 102
Pikt (AU 143 ml−1)3.3 × 102
Pikt (AU 286 ml−1)1.2 × 102
  • Initial inoculation level of D. bruxellensis 3 × 105 CFU ml−1. 143 and 286 AU correspond to 10- and 20-fold the amount toxin present in the supernatant, respectively.

In conclusion, we have here described two killer toxins that are active on Dekkera/Brettanomyces spoilage yeasts for the first time. Interestingly, Pikt and Kwkt maintain their killer activity in wine. Considering that Brettanomyces/Dekkera yeasts can cause heavy economic losses [9] and that most of the strategies used to control them to date are inadequate or detrimental to the final quality of the wine, Pikt and Kwkt may have great potential as biopreservative agents to be utilised during wine ageing to counteract undesired spoilage yeasts.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
View Abstract