The presence of yeasts in different kinds of table olivefermentation is common (Marquina et al., 1992; Llorente et al.,1997; Kotzekidou, 1997; Tassou et al., 2002; Durán Quintanaet al., 2003; Hernández et al., 2006). The predominant yeastspecies that have been isolated from Greek-style black olives areTorulaspora delbrueckii, Debaryomyces hansenii, and Crypto-coccus laurentii (Kotzekidou, 1997). Other workers (Marquinaet al., 1992) have isolated Pichia membranifaciens and relatedspecies as the dominant yeasts from spontaneous fermentationsof olive brines from Portugal as the dominant yeasts. In studiesperformed by Marquina et al. (1997) with olive brines from
⁎ Corresponding author. Tel.: +34 924 286200; fax: +34 924 286201.E-mail address: [email protected] (F. Pérez-Nevado).URL: http://eia.unex.es (F. Pérez-Nevado).
seven locations in Morocco, the most ubiquitous and abundantspecies were T. delbrueckii, Candida boidinii, and P. membra-nifaciens. Various studies have attributed to the yeast populationthe role of contributing to the sensorial characteristics of tableolives (Garrido et al., 1995; Sánchez et al., 2000). Since thesemicro-organisms could have an important effect on the quality ofthis product, their study could help in the selection of startercultures to be used in the elaboration of table olives. Yeasts suchas Candida krusei, and P. membranifaciens (originally Candidavalida), for example, are not considered spoilage yeasts andcould be used as starter cultures (Durán Quintana et al., 1979).
However, other surveys have associated the presence ofyeasts with different kinds of olive spoilage (Vaughn et al., 1969and 1972; Durán Quintana et al., 1979 and 1986). If film-forming yeasts are not controlled, they can rapidly oxidize thedesirable acidity in the storage brines of Sicilian-style andSpanish-type olives. Vaughn et al. (1972) found that Saccharo-
179A. Hernández et al. / International Journal of Food Microbiology 121 (2008) 178–188
myces cerevisiae (originally identified as S. oleaginosus), Sac-charomyces kluyveri, and Pichia anomala (originally identifiedas Hansenula anomala) can cause softening and gas-pocketformation in olives. Also, pink yeasts identified as Rhodotorulaglutinis, Rhodotorula minuta, and Rhodotorula rubra causeslow softening of olive tissue (Vaughn et al., 1969). The pre-sence of various yeast strains of S. cerevisiae and P. anomala hasbeen related to “alambrado” (bloater) spoilage in spontaneousfermentation in black olives (Durán Quintana et al., 1979). Otherspecies that have been related to this alteration are Pichia sub-pelliculosa (originally Hansenula subpelliculosa), Kluyvero-myces thermotolerans (originally Kluyveromyces veronae),Candida saitoana (originally Torulopsis candida), Candidanorvegica (originally Torulopsis norvegica), D. hansenii, andPichia fermentans.
Different methods have been used to control spoilage pro-duced by yeasts in table olives. The most commonly usedpractice in this industry of controlling the pH and salt level ofthe brine is insufficient to avoid these problems (Lamzira et al.,2005). Studies with an essential oil prepared from garlic had amajor effect in controlling the yeast population, but the orga-noleptic characteristics of the product were affected (Asehraouet al., 1997). Trials with pH adjusted to 4, added potassiumsorbate, and Lactobacillus plantarum inoculation have foundthat bloater spoilage was reduced drastically (Asehraou et al.,2002; Lamzira et al., 2005). The use of yeasts as starter culturecould protect the product against spoilage yeasts, selecting, for
example, killer yeasts which are known to be able to controlspoilage in the preservation of food. Killer yeasts can producetoxic proteins or glycoproteins (so-called killer toxins) that cancause death in other sensitive (killer-sensitive) yeast strains. Thekiller phenotype appears to be widely distributed within manyyeast genera (Schmitt and Breinig, 2002) some of them isolatedfrom a great variety of fermentation food processes (Llorente etal., 1997; Regodón et al., 1997; Gulbiniene et al., 2004). It isaffected by diverse ambient conditions like pH, includingtemperature, and the presence of salt (Woods and Bevan, 1968;Llorente et al., 1997; Marquina et al., 2001; Buzzini et al., 2004;Izgü and Altinbay, 2004). Moreover, its detection dependsstrongly on the sensitive strains used — the killing ability ofdifferent compounds may be underestimated or may evenremain unnoticed depending on the selection of the appropriatesensitive strain and other experimental conditions. For thisreason, the high variability of the killer phenomenon in natureprovides an exceptional tool for the discrimination of yeasts atthe strain level. Also, the use of these yeasts as a biocontrolmethod may improve table olives by reducing the requirementfor salt or such chemical preservatives as sorbic acid or similar.
The aim of the present work was to study the killer activity ofyeast strains isolated from seasoned green table olives. Sincesalt addition is a common practice in the production of tra-ditional fermented table olives, the effect of NaCl addition onthe killer expression was analyzed, as also was the effect of pH.In addition, the killer activity and sensitivity of isolates todifferent killer reference and wild yeast strains was analyzedwith a view to the potential use of these yeasts as biocontrolagents against spoilage yeasts.
2. Materials and methods
2.1. Yeast strains isolated from green table olives
Seventy-four indigenous yeast strains isolated from seasonedtable olive fermentation by Hernández et al. (2006) were tested.
Fig. 1. Percentage by genera of killer yeasts isolated from olive brines against the killer-sensitive reference strain EX73R at various pHs.
180 A. Hernández et al. / International Journal of Food Microbiology 121 (2008) 178–188
In prior studies, fifty-one of them had been pre-selected assuitable yeasts to be used as starter cultures in table olivefermentation (Table 1). In addition, 23 yeasts isolated from tableolives and classified as potential spoilage yeast species werestudied (Table 2).
Table 3Killer activity in diverse ranges of pH of yeasts isolated from olive brines againstEX73R killer-sensitive reference strain
Isolate Killeractivity(pH range)
Isolate Killeractivity(pH range)
Candida PichiaC. inconspicua FM47 – P. anomala FM5 7.5–8.0C. lusitaniae FM59 – P. anomala FM14 5.0–8.0C. maris FM15 6.0–8.0 P. anomala FM23 4.0–8.5C. maris FM46 – P. anomala FM25 –C. maris FM66 – P. anomala FM30 –C. maris FM67 – P. anomala FM31 3.5–8.5C. zeylanoides FM10 – P. anomala FM34 –C. zeylanoides FM71 3.5–5.0 P. anomala FM36 –Cryptococcus P. anomala FM37 –C. humicola FM29 3.5–8.0 P. anomala FM38 –C. humicola MP11 – P. anomala FM40 3.5–5.0Debaryomyces P. anomala FM52 –D. hansenii FM1 3.5–6.0 P. anomala FM58 3.5–8.5D. hansenii FM32 – P. anomala FM62 –D. hansenii FM72 3.5–8.5 P. anomala FM73 –Kluyveromyces P. anomala FM74 –K. marxianus FM2 – P. anomala FM79 3.5–8.5K. marxianus FM6 – P. guilliermondii MP6 3.5–8.5K. marxianus FM7 – SaccharomycesK. marxianus FM11 3.5–8.0 S. cerevisiae FM8 3.5–8.5K. marxianus FM12 3.5–8.0 S. cerevisiae FM9 –K. marxianus FM19 4.5–7.0 S. cerevisiae FM20 6.0–7.0K. marxianus FM24 – S. cerevisiae FM39 –K. marxianus FM26 3.5–7.0 S. cerevisiae FM41 3.5–5.5K. marxianus FM27 3.5–8.5 S. cerevisiae FM43 3.5–8.5K. marxianus FM69 – S. cerevisiae FM44 3.5–8.5K. marxianus FM78 – S. cerevisiae FM51 –
S. cerevisiae FM68 –
2.2. Reference strains
For the killer activity screening, we used several referencestrains for killer and sensitive traits. These included four killer-sensitive strains: EX33 and EX73R (S. cerevisiae industrial winestrains, isolated and selected by Regodón et al., 1997); 5×47(killer-sensitive strain, provided by Dr R.B. Wickner, Section ofGenetics of Simple Eukaryotes, National Institute of Diabetes,Digestive and Kidney Diseases, NIH Bethesda, MD 20892,USA), and MNN9 (Laboratory collection strain, provided byDrs Luis M. Hernández and Isabel Olivero, MicrobiologyDepartment, University of Extremadura, Spain). Five killeryeasts were used: K1 (JCR2 killer strain, provided by Dr R.B.Wickner); K2 (EX85 killer, S. cerevisiae industrial wine strainisolated and selected by Regodón et al., 1997); K28 (GY-2-3akiller, obtained from Manfred Schmitt, Angewandte Moleku-larbiologie, Universität des Saarlandes, Im Sadtwadl, Gebäude2, D-66123 Saarbrücken, Germany); KL (Kluyveromyces lactis,wild type provided by Dr Ángel Domínguez, Universidad deSalamanca); and HM1 (HM22 killer strain of Williopsis sa-turnus, provided by Dr J.C. Ribas, Universidad de Salamanca).
2.3. Assays of the killer phenotype at different pH’s and NaClconcentrations
These tests were performed as described by Pérez-Nevadoet al. (2006). Plates of YEPD-MB medium (0.5% yeast extract,1% peptone, 2% glucose, 2% agar, 0.003% methylene blue,0.1 M sodium citrate) with different pH values (pH 3.5, 4.0, 4.5,5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, and 8.5) were seeded with thekiller-sensitive reference strain EX73R pre-grown for 48 h onYEPD-agar slants. In order to analyze the influence of NaCl onthe killer phenotype, YEPD-MB adjusted to pH 4.0 andsupplemented or not with various NaCl concentrations (5, 8,and 10%) was seeded with the EX73R killer-sensitive strain.Strains to be tested for killer activity in these media were loadedonto the seeded agar. Colonies exhibiting clear halos on the
Fig. 2. Clear zones of inhibition of killer yeasts in a lawn of a killer reference strain in media with two different NaCl concentrations (5% and 8%, w/v).
181A. Hernández et al. / International Journal of Food Microbiology 121 (2008) 178–188
sensitive lawns after 3–7 days of incubation at 20 °C wereconsidered to be killer positive.
To test for killer-sensitive traits, YEPD-MB pH 4.0 plateswithout added NaCl were seeded with the isolates to be testedand overlaid with K1, K2, K28, KL, and HM1 killer referencestrains. Isolates exhibiting clear halos on the plates were con-sidered to be killer-sensitive towards the respective referencekiller strains.
2.4. Cross-reaction trials between different yeasts isolated fromtable olives
Studies were performed to analyze cross-reactions betweenyeasts that can cause spoilage in table olives and pre-selectedyeasts to be used as a starter culture. On the surface of YEPD-
Fig. 3. Percentage of killer yeasts in each of the yeast genera studied in
MB pH 4.0, the spoilage yeasts isolated from table olives werespread as a lawn. Then, isolates were streaked on top of theseeded agar. Colonies exhibiting clear halos on the spoilagelawns after 3–7 days of incubation at 20 °C were considered tobe killer positive.
3. Results and discussion
3.1. Killer yeasts at different pH
The killer activity of the yeast isolates was analyzed at differentpH's (3.5 to 8.5) against the killer-sensitive strain EX73R.At all thepH's analyzed, the percentage of killer yeasts was always greaterthan 19%. There was no great variation in killer percentages at pH3.5 to 8.0, the values remaining between 33.3 and 39.2%. The
media with different NaCl concentrations (0, 5, 8, and 10%, w/v).
Table 4Killer activity in yeasts isolates from olive brines against four sensitive reference strains and sensitivity of isolated against five killer reference strains
Isolate Sensitive strain Killer reference strain
MNN9 5x47 EX33 EX73R K1 K2 K28 KL HM1
CandidaC. inconspicua FM47 – – – – S S – – SC. lusitaniae FM59 – – – – S S – – SC. maris FM15 – K – – – – – – SC. maris FM46 – – – – S S – – SC. maris FM66 K – – – S S S – SC. maris FM67 – – – – S S – – SC. zeylanoides FM10 – – – – S S – – SC. zeylanoides FM71 – K K K – – – – S
CryptococcusC. humicola FM29 K K K K – – – – SC. humicola MP11 – – – – S S – – S
DebaryomycesD. hansenii FM1 – K K K S S – – SD. hansenii FM32 – – – – S S S – SD. hansenii FM72 K K K K – – – – S
KluyveromycesK. marxianus FM2 K – K K – – – – SK. marxianus FM6 – – – – S S – – SK. marxianus FM7 – – – – – S – – –K. marxianus FM11 – K K K – – – S SK. marxianus FM12 – K K K – – – – SK. marxianus FM19 K – – – – – – – SK. marxianus FM24 – – – – S S – – SK. marxianus FM26 K K K K – – – – SK. marxianus FM27 K K K K – – – – SK. marxianus FM69 – – – – S S – – SK. marxianus FM78 – – – – S S S – S
PichiaP. anomala FM5 K – – – S S – – SP. anomala FM14 – K – – – – – S SP. anomala FM23 K K K K – – – – SP. anomala FM25 – – – – S S – – SP. anomala FM30 – – – – – – – – SP. anomala FM31 – K K K – – – – SP. anomala FM34 – – – – S S S – SP. anomala FM36 – – – – S S S – SP. anomala FM37 – – – – S S S – SP. anomala FM38 – – – – S S – – SP. anomala FM40 K K K K – – – – SP. anomala FM52 – – – – S S – – SP. anomala FM58 K K K K – – – – SP. anomala FM62 – – – – S S – – SP. anomala FM73 – – – – S S – – SP. anomala FM74 – – – – S S – – SP. anomala FM79 K K K K – S – – SP. guilliermondii MP6 – K K K – – – S S
SaccharomycesS. cerevisiae FM8 – K K K – – – – SS. cerevisiae FM9 – – – – – – – – –S. cerevisiae FM20 – – – – – – – – –S. cerevisiae FM39 – – – – S S – – SS. cerevisiae FM41 K K K K – – – – SS. cerevisiae FM43 K K K K – – – – SS. cerevisiae FM44 K – – K – – – – S
182 A. Hernández et al. / International Journal of Food Microbiology 121 (2008) 178–188
Table 4 (continued)
Isolate Sensitive strain Killer reference strain
MNN9 5x47 EX33 EX73R K1 K2 K28 KL HM1
SaccharomycesS. cerevisiae FM51 – – – – S S – – SS. cerevisiae FM68 – – – – S S – – S
K: killer activity against the sensitive strain; S: sensitivity against killer reference strain; –: neutral activity.
183A. Hernández et al. / International Journal of Food Microbiology 121 (2008) 178–188
highest percentage of killer yeasts was obtained at pH 5 (39.2%) sothat this pH was considered to be the best at which to analyze theactivity of our yeasts. In contrast, pH 8.5, the highest tested,corresponded to the lowest killer percentage (19.6%). The resultsreported by Marquina et al. (1997) studying olive brines fromMorocco were different: only 12.3% of their yeasts isolated fromolive brines were killer at pH 4.8.
Fig. 1 shows the percentage of killer yeasts in each genus atthe different pH's tested. In all the genera, killer yeasts werefound against one or more of the killer-sensitive yeasts used.Candida was the genus with the lowest percentages of killerstrains at the different pH's tested (always less than or equal to20%). Debaryomyces was the genus with the highest percen-tages of killer (66.7%), at least at low pH (between 3.5 and 6.0).In most genera, the killer percentage diminished at the highestpH, and indeed, no strain of Candida was killer at pH 8.5.
Other studies have found killer yeasts belonging to verymany genera and species. In studies performed with yeastsisolated from spontaneous fermentations of olive brines, strainsof the genera Pichia, Kluyveromyces, Candida, and Torulas-pora were killer against killer-sensitive strains of S. cerevisiae(Marquina et al., 1992 and 1997). Aguiar and Lucas (2000)working with yeasts from different sources, found killer yeastsin various genera including Saccharomyces, Candida, Pichia,Rhodotorula, Torulaspora, Fellomyces, and Zygosaccharo-myces. At wineries, Zagorc et al. (2001) isolated killer yeastsbelonging to P. anomala, Pichia kluyveri, Hanseniasporauvarum, and Candida rugosa; whereas Gutiérrez et al. (2001)only found killer yeasts belonging to the Saccharomyces genus.
We found that each killer strain was able to produce a toxinactive over different ranges of pH (Table 3). The variety ofpatterns one observes in the table suggest that the toxins could bebiochemically distinct. Most killer strains presented a broad pHactivity range, between 3.5 or 4.0 and 8.0 or 8.5; and all thegenera tested had strains with a broad killer activity. Otherauthors have observed the relationship between killer activityand pH (Woods and Bevan, 1968, were the first), although moststudies performed with killer yeasts have found that killer toxinsgenerally are stable only over a narrow pH range. There is a well-defined optimal pH for the killer activity of many killer toxins.Most have an optimal pH between 4.1 and 4.7 (Chen et al., 2000;Soares and Sato, 2000; Izgü and Altinbay, 2004). Marquina et al.(2001) observed that killer toxin isolated from D. hansenii hadan optimal stability at pH 4.5, but was destroyed completely atpH 5.1. Similar results were obtained for killer toxins fromDebaryomyces occidentalis (originally identified as Schwan-niomyces occidentalis), Kluyveromyces phaffi, and Pichia(Chen et al., 2000; Ciani and Fatichenti, 2001; Ceccato-Antoniniet al., 2004; Izgü and Altinbay, 2004). However, other workers
have found killer toxins with a broad pH stability range (from 2.0to 11.0) in strains ofW. saturnus var. mrakii (Ashida et al., 1983)or Williopsis saturnus var. saturnus (Ohta et al., 1984;Komiyama et al., 1995). Buzzini et al. (2004) report that akiller protein produced by a yeast isolated from tropical habitats,Williopsis saturnus, is active over a broad range of pH (4.5-8.0).
3.2. Killer activity in the presence of various NaClconcentrations
The effect of the presence of salt on killer activity wasstudied by testing all the isolates against a killer-sensitive strain(EX73R). At NaCl concentrations of 5 and 8%, the percentageof killer was greater than without added salt: 49.1 and 50.9% ofyeasts were killer at 5 and 8% of salt, respectively, whereaswithout added salt only 37.3% were killer. Differences betweenthe clear zones of killer yeast inhibition were found at differentNaCl concentrations. At higher salt concentrations, the clearzones were larger (Fig. 2). However, the percentage of killerstrains diminished (to 28.3%) at 10% NaCl. Thus, high pH andsalt concentrations can affect the killer strain, killer toxin pro-duction, or killer toxin activity negatively. These results couldindicate that the absence of salt in the medium might haveresulted in an underestimate of killer yeast numbers.
Fig. 3 shows the killer activity results by genus. Except forDebaryomyces, the percentage of killer strains increased at 5 and8% salt concentration, but decreased at 10%. For the genus De-baryomyces, the killer percentage was the same with and withoutsalt added.We also observed that the killer activity at different NaClconcentrations was strain dependent (data not shown).
Marquina et al. (1997) also found that killer characterdepended on the salt concentration used in the trial with thepercentage of killer strains increasing at 3 or 6% NaCl. Ourresults are concordant with those of Llorente et al. (1997) whofound that killer activity was dependent on both the killer strainand the sensitive strain, but in most cases the apparent toxicitywas significantly enhanced as the salt concentration was in-creased. Also, Aguiar and Lucas (2000) found an abrupt de-crease in killer phenotype expression at high salt concentrations(above 11.7% NaCl). Those authors suggest the existence of aphenotype relationship between high halotolerance and killercapacity in the presence of high salt concentrations.
3.3. Intraspecific differences in killer susceptibility and activity
Table 4 presents the results for the interaction between referencekiller yeasts and yeast isolates at pH 4.0, a pH near that of olivesduring fermentation. The killer activity and resistance of selectedstrains isolated from olive brines were determined by interaction
Table 5Interactions between yeast strains isolated from table olives and yeasts that cause spoilage
C. inconspicua FM47 – – – – – – – – – – K –C. lusitaniae FM59 – – – – – – – – – – – –C. maris FM15 – – – – – K – – K – K KC. maris FM46 – – – – – – – – – – – –C. maris FM66 – – – K K – – – – K – –C. maris FM67 – – – – – – – – – – – –C. zeylanoides FM10 – – – – – – – – – – – –C. zeylanoides FM71 – – – – K K – – – K – –C. humicola FM29 K – K – K K – – – K – KC. humicola MP11 – – – – – – – – – – – –D. hansenii FM1 K – – – K – – – K – – –D. hansenii FM32 – – – – – – – – – – – –D. hansenii FM72 K K K K K K – – K K – KK. marxianus FM2 – – – – – – – – – – – –K. marxianus FM6 – – – – – – – – – – – –K. marxianus FM7 – – – – – – – – – – – –K. marxianus FM11 – – – – – K – – K – K KK. marxianus FM12 K – K – K K – – K – K KK. marxianus FM19 K K K K K K – K K K K KK. marxianus FM24 – – – – – – – – – – – –K. marxianus FM26 – K K K K K – K K K K KK. marxianus FM27 K – K K K K K K K K K KK. marxianus FM69 – – – – – – – – – – – –K. marxianus FM78 – – – – – – – – – – – –P. anomala FM5 – – – – – – – – – – – –P. anomala FM14 – – K – – K – – K – K KP. anomala FM23 K K K – K K – K K K K KP. anomala FM25 – – – – – – – – – – – –P. anomala FM30 – – – – – – – – – – – –P. anomala FM31 – K K K K K – K K K K KP. anomala FM34 – – – – – – – – – – – –P. anomala FM36 – – – – – – – – K – – –P. anomala FM37 – – – – – – – – – – – –P. anomala FM38 – – – – – – – – – – – –P. anomala FM40 – – – – – K – – – – – –P. anomala FM52 – – – – – – – – – – – –P. anomala FM58 – K – – K K – K K – K –P. anomala FM62 – – – – – – – – – – – –P. anomala FM73 – – – – – – – – – – – –P. anomala FM74 – – – – – – – – – K – –P. anomala FM79 K – K – K K K K K K K KP. guilliermondii MP6 – K K – K K – – K K K KS. cerevisiae FM8 K – – – K K – – K – K KS. cerevisiae FM9 – – – – – – – – – – – –S. cerevisiae FM20 – – – – – – – – – – – –S. cerevisiae FM39 – – – – – K – – – – – –S. cerevisiae FM41 – K K – K K K – – – – –S. cerevisiae FM43 – – – – K K – – K K – –S. cerevisiae FM44 K – K K – K – – K – – –S. cerevisiae FM51 – – – – – – – – – – – –S. cerevisiae FM68 – – – – – – – – – – – –
K: killer activity against the spoilage strain; –: neutral activity.
184 A. Hernández et al. / International Journal of Food Microbiology 121 (2008) 178–188
with killer and killer-sensitive reference strains.Most of the isolateswere killed by one or more killer strains (96.1%). Only two isolates(3.9%), S. cerevisiae FM9 and S. cerevisiae FM20, were resistant
to all five killer reference strains. These two strains were not killeragainst any of the four reference sensitive strains used. They weretherefore considered killer-neutral. Nearly half the isolates (47.1%)
– – K K – K – – – K – 5 (21.7)– – – K – – – – – – – 1 (4.3)K – – – – – – – – – – 5 (21.7)– – – – – – – – – – – 0 (0)– K K K K K K K K – – 11 (47.8)– – – – – – – – – – – 0 (0)– – – – – – – – K – – 1 (4.3)– – – – – K – – K – – 5 (21.7)– K K – – K K – – – – 10 (43.5)– K – – – K – – K – – 3 (13)– K – K – K – – K – – 7 (30.4)– – – – – – – – K – – 1 (4.3)– – K – K K K K K K – 16 (69.6)– – – – – – – – – – – 0 (0)– – – K – – – – – – – 1 (4.3)– – – K – K – – K – – 3 (13.0)– – – – – K – – K K – 7 (30.4)– – – K – – – – K K – 10 (43.5)– K K – K K K K K K – 19 (82.6)– – – – – – – – – – – 0 (0)K K K – K K K K K K – 19 (82.6)K K K – K K K K K K – 20 (87.0)– – – – – – – – – – – 0 (0)– – – – – – – – – – – 0 (0)– – – K – – – – K – – 2 (8.7)– – – K – K – – K – – 8 (34.8)– K K – – K K K K K – 17 (73.9)– – – – – – – – – – – 0 (0)K – – – – – – – – – – 1 (4.3)– K K – – K K K K K – 17 (73.9)– – – – – – – – K – – 1 (4.3)– – – – – – – – K – – 2 (8.7)– – – – – – – – – – – 0 (0)– – – – – – – – – – – 0 (0)– – – K – K – – – – – 3 (13.0)– – – – – – – – – – – 0 (0)– – K – K K – – K K – 11 (47.8)– – K – – – – – – – – 1 (4.3)– – – – – – – – – – – 0 (0)– – – – – – – – K – – 2 (8.7)K K K – – K K K K K – 18 (78.3)K K – K – K – K K K – 15 (65.2)K K – K – K – – K K K 13 (56.5)– K – – – – – – K – – 2 (8.7)– – – – – – – – – – – 0 (0)– – – K – – – – – – – 2 (8.7)– – – K – K K – – K – 9 (39.1)K – – – – K K – – K – 8 (34.8)– – – – – K – – K K – 8 (34.8)– – – – – – – – K – – 1 (4.3)– – – – – – – – – – – 0 (0)
185A. Hernández et al. / International Journal of Food Microbiology 121 (2008) 178–188
were considered killer-sensitive strains because they were killed byone or more killer reference strains and were not able to kill any ofthe reference sensitive strains.
The most sensitive strains which only were resistant to theKL killer reference strain, were Candida maris FM66, D.hansenii FM32, Kluyveromyces marxianus FM78, P. anomala
186 A. Hernández et al. / International Journal of Food Microbiology 121 (2008) 178–188
FM34, P. anomala FM36, and P. anomala FM37. The killerreference strain HM1 was the most active towards the isolates.Only three strains (K. marxianus FM7, S. cerevisiae FM9, andS. cerevisiae FM20) were able to resist the killer toxin of thisyeast. Some of the strains (35.3%) were killed only by the HM1reference strain. The other isolates were killed by two or morekiller strains (58.8%), and one isolate (K. marxianus FM7) waskiller-sensitive only to killer reference K2. Only three isolateswere sensitive to the K. lactis killer reference strain. These wereK. marxianus FM11, P. anomala FM14, and Pichia guillier-mondii MP6.
Intraspecific differences were found in killer performance.All the species studied exhibited varying killer-sensitivitypatterns (Table 4): 5 patterns in Candida, 3 in Debaryomyces,8 in Kluyveromyces, 9 in Pichia, and 5 in Saccharomyces.Similar findings were reported by Yap et al. (2000) workingwith yeast species representative of wine fermentation micro-flora. However, in that work S. cerevisiae strains had a singlekiller-sensitivity pattern, whereas in the present study there werethree different patterns. Vaughan-Martini et al. (1996) andSangorrín et al. (2002) indicate that the sensitivity of yeasts todifferent killer toxins is a strain-related property, and, on thebasis of this variability, several authors (Vaughan-Martini et al.,1996; Buzzini and Martini, 2000; Sangorrín et al., 2002) havesuggested that the killer-sensitivity relationships can be used tofingerprint yeasts at the strain level.
3.4. Cross-reaction trials between isolates and spoilage yeasts
The interaction between isolates and spoilage yeasts maydetermine the effect of a yeast starter culture on other yeasts thatcan cause spoilage. In Table 5, one observes that most of theisolates (74.5%) were killer against one or more spoilagestrains. In studies performed by Marquina et al. (1997) withyeasts isolated from olive brines, 42.1% of the yeasts were killeragainst one or more of the yeast isolates.
As expected, all the yeast species studied exhibited killerabilities, although there were intraspecific differences. Thisseems to suggest that, as was noted above, the toxins of eachstrain could be biochemically distinct. These results are inagreement with those of Marquina et al. (1997) who found thatthe detection of killer yeasts depends strongly on the sensitivestrain used. Also, other studies have found that sensitive inter-actions are more frequent among yeasts isolated from differenthabitats, i.e., killer yeasts are more effective against foreign thanindigenous species (Yap et al., 2000; Carreiro et al., 2002;Trindade et al., 2002).
Nine isolates were considered of potential interest becausethey killed most of the spoilage yeasts analyzed: 1 D. hansenii(FM72), 3 K. marxianus (FM19, FM26 and FM27), 3 P.anomala (FM23, FM31 and FM79), 1 P. guilliermondii (MP6),and 1 S. cerevisiae (FM8). These yeasts could be used as bio-control starters against different spoilage yeasts. In contrast,some yeasts (C. maris FM46 and FM67; K. marxianus FM2,FM24, FM69, and FM78; P. anomala FM25, FM37, FM38,FM52, and FM73; and S. cerevisiae FM20 and FM68) were notkiller towards any spoilage yeast. Finally, Trichosporon cuta-
neum FM13, was the spoilage yeast most sensitive to the iso-lates: 27 isolates were able to kill this spoilage yeasts. Otherspoilage yeasts affected by many isolates were R. glutinis MP9and C. rugosa FM18, which were killed by 23 and 21 isolates,respectively. In contrast, T. cutaneum FM54 and Cryptococcusalbidus FM22 were the least affected spoilage yeasts. Only 1and 3 isolates killed them, respectively.
There have been various studies of the killer cross-reactionsbetween yeasts of diverse genera. In the survey by Marquinaet al. (1992), all strains of P. anomala isolated from olive brineswere killer, being able to kill Candida, Debaryomyces, Kluy-veromyces, Pichia, and Saccharomyces yeasts from a collectionculture. In studies performed with wine yeasts, strains of P.anomala, Kluyveromyces wickerhamii, and K. phaffi were killeragainst other yeast genera, including Dekkera/Brettanomyces,and Hanseniaspora uvarum (Ciani and Fatichenti, 2001;Comitini et al., 2004). Yap et al. (2000) found that strains ofP. anomala, W. saturnus, and K. lactis isolated from differentsources killed a broad range of yeast strains. Buzzini et al.(2004) have shown the potential of killer toxin secreted by otheryeasts such as Williopsis saturnus to control the growth ofpathogenic yeast strains belonging to C. glabrata, Issatchenkiaorientalis, and P. guilliermondii. In general, non-Saccharo-myces species, such as Zygosaccharomyces bailii and H. uva-rum, show a broad-spectrum antimycotic potential, being ableto kill yeasts and fungi belonging to a great variety of generaincluding Candida, Sporothrix, Heterobasidium, Postia, andSerpula (reviewed by Schmitt and Breinig, 2002). In contrast,killer toxins produced by S. cerevisiae are generally active onlyagainst yeasts of this genus (Ciani and Fatichenti, 2001), al-though other workers (Soares and Sato, 2000; Zagorc et al.,2001) have described killer toxins produced by S. cerevisiaethat are active against C. glabrata, P. anomala, P. kluyveri,Pichia pijperi, and C. rugosa.
Comparing these results with those of the killer activityagainst sensitive reference strains, one observes different typesof behaviour. Some isolates killed one or more spoilage yeasts,but could not kill any of the sensitive reference strains. This wasthe case of Cryptococcus humicola MP11, Candida incon-spicua FM47, Candida lusitaniae (anamorph of Clavisporalusitaniae) FM59, Candida zeylanoides FM10, D. hanseniiFM32, K. marxianus FM6 and FM7, P. anomala FM30, FM34,FM36, FM62 and FM74, and S. cerevisiae FM9, FM39 andFM51. However, other yeasts (C. maris FM46 and FM67; K.marxianus FM24, FM69, and FM78; P. anomala FM25, FM37,FM38, FM52, and FM73 and S. cerevisiae FM68) were able tokill neither sensitive reference strains nor spoilage yeasts. In thissense, one can conclude that the variability in the patterns ofkiller activity against different strains shows that the productionof killer toxins is not a characteristic of the genus or species, butis expressed by each particular strain. This is in agreement withthe findings of other workers such as Carreiro et al. (2002).
4. Conclusions
We found a great variability in the killer activity and sus-ceptibility of the isolated yeast strains. In general, killer activity
187A. Hernández et al. / International Journal of Food Microbiology 121 (2008) 178–188
was present at all the pH's and NaCl concentrations tested, butdecreased at the highest pH (8.5) and NaCl concentration(10%). This behaviour was highly strain dependent, with eachstrain showing killer activity over different ranges of pH andsalt concentration. All the results confirmed the highly poly-morphic expression of the killer activity. The killer activityunder different conditions could thus be used for the simple andrapid characterization of yeast strains of industrial interest. Incross-reactions with spoilage yeasts, several isolates showed abroad killer spectrum towards yeasts that could cause spoilage.This suggests that starter cultures of these isolates could be usedas a biocontrol method in olive brine fermentation, acting ascompetitors with spoilage yeasts. This could improve the end-product by reducing the need for different chemical preserva-tives, including reducing the concentrations. Further studieswould be required to determine the nature of the toxin and thekiller genetics.
Acknowledgments
The authors would like to thank Drs Manuel Ramírez Fer-nández, Luis M. Hernández, and Isabel Olivero (Dpto. de Mi-crobiología, Universidad de Extremadura, Spain) forsupplying killer and killer-sensitive reference yeast strains.A. Hernández was the beneficiary of a predoctoral grant fromthe Consejería de Educación y Tecnología (Junta de Extrema-dura, Spain).
References
Aguiar, C., Lucas, C., 2000. Yeasts killer/sensitivity phenotypes and haloto-lerance. Food Technology and Biotechnology 38, 39–46.
Asehraou, A., Mohieddine, S., Faid, M., Serhrouchni, M., 1997. Use of anti-fungal principles from garlic for the inhibition of yeasts and moulds infermenting green olives. Grasas y Aceites 48, 68–73.
Asehraou, A., Peres, C., Faid, M., Brito, D., 2002. Reducing the bloater spoilageincidence in fermented green olives during storage. Grasas y Aceites 53,330–334.
Ashida, S., Shimazaki, T., Kitano, K., Hara, S., 1983. New killer toxin ofHansenula mrakii. Agricutural and Biological Chemistry 47, 2953–2955.
Buzzini, P., Corazzi, L., Turchetti, B., Buratta, M., Martini, A., 2004. Charac-terization of the in vitro antimycotic activity of a novel killer protein fromWilliopsis saturnus DBVPG-4561 against emerging pathogenic yeasts.FEMS Microbiology Letters 238, 359–365.
Buzzini, P., Martini, A., 2000. Differential growth inhibition as a tool to increasethe discriminating power of killer toxin sensitivity in fingerprinting ofyeasts. FEMS Microbiology Letters 193, 31–36.
Carreiro, S.C., Pagnocca, F.C., Bacci, M., Bueno, O.C., Hebling, M.J.A.,Middelhoven, W.J., 2002. Occurrence of killer yeasts in leaf-cutting antnests. Folia Microbiologica 47, 259–262.
Ceccato-Antonini, S.R., Tosta, C.D., da Silva, A.C., 2004. Determination ofyeast killer activity in fermenting sugarcane juice using selected ethanol-making strains. Brazilian Archives of Biology and Technology 47,13–23.
Chen, W.-B., Han, Y.-F., Jong, S.-C., Chang, S.-C., 2000. Isolation, purificationand characterization of a killer protein from Schwanniomyces occidentalis.Applied and Environmental Microbiology 66, 5348–5352.
Ciani, M., Fatichenti, F., 2001. Killer toxin of Kluyveromyces phaffii DBVPG6076 as a biopreservative agent to control apiculate wine yeasts. Applied andEnvironmental Microbiology 67, 3058–3063.
Comitini, F., De Ingeniis, J., Pepe, L., Mannazzu, I., Ciani, M., 2004. Pichiaanomala and Kluyveromyces wickerhamii killer toxins as new tools against
Durán Quintana, M.C., González, F., Garrido, A., 1979. Aceitunas negras alnatural en salmuera. Ensayos de producción de alambrado. Inoculación dediversos microorganismos aislados de salmueras de fermentación. Grasas yAceites 30, 361–367.
Durán Quintana, M.C., García, P., Garrido, A., 1986. Fermentación en medioaeróbico de aceitunas maduras en salmuera con inyección alternante de aire.Estudio de la influencia de la adición de cloruro cálcico sobre la textura.Grasas y Aceites 37, 242–249.
Durán Quintana, M.C., García, P., Garrido, A., 2003. Características delcrecimiento de levaduras de aceitunas de mesa a bajas temperaturas. Grasasy Aceites 54, 264–271.
Garrido, A., García, P., Brenes, M., 1995. Olive fermentations. In: Rem, H.-J.,Reed, G. (Eds.), Biotechnology. VCH, Weinheim, pp. 593–627.
Gulbiniene, G., Kondratiene, L., Jokantaite, T., Serviene, E., Melvydas, V.,Petkuniene, G., 2004. Occurrence of killer yeast strains in fruit and berrywine yeast populations. Food Technology and Biotechnology 42, 159–163.
Gutiérrez, A.R., Epifanio, G., Garito, P., López, R., Santamaría, P., 2001. Killeryeast: incidence in the ecology of spontaneous fermentation. AmericanJournal of Enology and Viticulture 52, 352–356.
Hernández, A., Martín, A., Aranda, E., Pérez-Nevado, F., Córdoba, M.G., 2006.Identification and characterization of yeast isolated from the elaboration ofseasoned green table olives. Food Microbiology 24, 346–351.
Izgü, F., Altinbay, D., 2004. Isolation and characterization of the K5-type yeastkiller protein and its homology with an exo-beta-1,3-glucanase. BioscienceBiotechnology and Biochemistry 68, 685–693.
Kotzekidou, P., 1997. Identification of yeasts from black olives in rapid systemmicrotitre plates. Food Microbiology 14, 609–616.
Lamzira, Z., Asehraou, A., Brito, D., Oliveira, M., Faid, M., Peres, C., 2005.Reducing the bloater spoilage during lactic fermentation of Moroccan greenolives. Food Technology and Biotechnology 43, 343–377.
Llorente, P., Marquina, D., Santos, A., Peinado, J.M., Spencer-Martins, I., 1997.Effect of salt on the killer phenotype of yeasts from olive brines. Applied andEnvironmental Microbiology 63, 1165–1167.
Marquina, D., Barroso, J., Santos, A., Peinado, J.M., 2001. Short communica-tion — Production and characteristics of Debaryomyces hansenii killertoxin. Microbiological Research 156, 387–391.
Marquina, D., Peres, C., Caldas, F.V., Marques, J.F., Peinado, J.M., Spencer-Martins, I., 1992. Characterization of the yeast population in olive brines.Letters in Applied Microbiology 14, 279–283.
Marquina, D., Toufani, S., Llorente, P., Santos, A., Peinado, J.M., 1997. Killeractivity in yeast isolates from olive brines. Advances in Food Science 19,41–46.
Ohta, Y., Tsukada, Y., Sugimori, T., 1984. Production, purification andcharacterization of Hyi, an anti-yeast substance, produced by Hansenula–Saturnus. Agricultural and Biological Chemistry 48, 903–908.
Pérez-Nevado, F., Albergaria, H., Hogg, T., Girio, F., 2006. Cellular death oftwo non-Saccharomyces wine-related yeasts during mixed fermentationswith Saccharomyces cerevisiae. International Journal of Food Microbiology108, 336–345.
Regodón, J.A., Pérez, F., Valdés, M.E., De Miguel, C., Ramírez, M., 1997. Asimple and effective procedure for selection of wine yeast strains. FoodMicrobiology 14, 247–254.
Sánchez, A.H., De Castro, A., Rejano, L., Montano, A., 2000. Comparativestudy on chemical changes in olive juice and brine during green olivefermentation. Journal of Agricultural and Food Chemistry 48, 5975–5980.
Sangorrín, M., Zajonskovsky, I., Van Broock, M., Caballero, A., 2002. The useof killer biotyping in an ecological survey of yeast in an old Patagonianwinery. World Journal of Microbiology and Biotechnology 18, 115–120.
Schmitt, M.J., Breinig, F., 2002. The viral killer system in yeast: from molecularbiology to application. FEMS Microbiology Reviews 26, 257–276.
Soares, G.A., Sato, H.H., 2000. Characterization of the Saccharomycescerevisiae Y500-4L killer toxin. Brazilian Journal of Microbiology 31,291–297.
188 A. Hernández et al. / International Journal of Food Microbiology 121 (2008) 178–188
Tassou, C.C., Panagou, E.Z., Katsaboxakis, K.Z., 2002. Microbiological andphysicochemical changes of naturally black olives fermented at differenttemperatures and NaCl levels in the brines. Food Microbiology 19,605–615.
Trindade, R.C., Resende, M.A., Silva, C.M., Rosa, C.A., 2002. Yeastsassociated with fresh and frozen pulps of Brazilian tropical fruits. Systematicand Applied Microbiology 25, 294–300.
Vaughan-Martini, A., Cardinalli, G., Martini, A., 1996. Differential killersensitivity as a tool for fingerprinting wine-yeast strains of Saccharomycescerevisiae. Journal of Industrial Microbiology 17, 124–127.
Vaughn, R.H., Jakubczyk, K.T., McMillan, J.D., Higinio Thomas, E., Dave, B.A., Crampton, V.M., 1969. Some pink yeast associated with softening ofolives. Applied Microbiology 18, 771–775.
Vaughn, R.H., Stevenson, K.E., Dave, B.A., Park, H.C., 1972. Fermenting yeastassociated with softening and gas-pocket formation in olives. AppliedMicrobiology 23, 316–320.
Woods, D.R., Bevan, E.A., 1968. Studies on nature of killer factor produced bySaccharomyces cerevisiae. Journal of General Microbiology 51, 115–126.
Yap, N.A., Lopes, M.D., Langridge, P., Henschke, P.A., 2000. The incidence ofkiller activity of non-Saccharomyces yeasts towards indigenous yeastspecies of grape must: potential application in wine fermentation. Journal ofApplied Microbiology 89, 381–389.
Zagorc, T., Maraz, A., Cadez, N., Jemec, K.P., Peter, G., Resnik, M., Nemanic,J., Raspor, P., 2001. Indigenous wine killer yeasts and their application as astarter culture in wine fermentation. Food Microbiology 18, 441–451.
