Killer Yeasts: Incidence in the Ecology of Spontaneous Fermentation

Transcription

1 352 Gutiérrez et l. Killer Yests: Incidence in the Ecology of Spontneous Fermenttion A.R. Gutiérrez, 1 * S. Epifnio, 2 P. Grijo, 2 R. López, 2 nd P. Sntmrí 2 The occurrence of the killer chrcter ws studied in 1320 yests isolted from 12 wineries of the Rioj DOC, Spin, t different stges of spontneous fermenttion, for both white nd red wines. The objective of the study ws to investigte the importnce of the killer phenotype in vinifiction t different wineries. Killer, neutrl, nd sensitive strins were found mong the Scchromyces cerevisie isoltes, while the killer phenotype ws bsent from the non-scchromyces cerevisie strins. The distribution of the killer chrcter ws linked to the production re (Rioj Alt, Rioj Bj, nd Rioj Alves) nd to the production technology used (vinifiction with destemming/crushing nd crbonic mcertion). The only killer strins found were in the Rioj Bj. In the crbonic mcertion vinifictions, lmost 100% of the strins were sensitive. Monitoring of this killer chrcter ws crried out over five yers t the winery of the Centre for Reserch nd Agriculturl Development of L Rioj with the following results: Scchromyces cerevisie strins with different phenotypes coexist t the different fermenttive stges, nd the dominnt phenotype vries with the yer. In ddition, geneticlly identicl strins were detected with different killer phenotypes, perhps indicting tht the killer chrcter is of lower technologicl importnce thn hs been previously supposed. Key words: Killer, Scchromyces cerevisie, spontneous fermenttion, crbonic mcertion The killer chrcter ws discovered in 1963 by Bevn nd Mkower [4] nd hs since been described for severl species nd gener of wine yests [19,24,25]. Three different phenotypes re found with respect to this fctor: killer (K+R+), sensitive (K-R-), nd neutrl (K-R+). The strins with killer phenotype produce toxin tht is lethl to the sensitive ones; neither the sensitive strins nor the neutrl ones produce the toxin, but the neutrls re resistnt to it. The cpcity of the Scchromyces cerevisie to produce nd to resist the toxin is determined geneticlly by virl-type prticles (RNA2c) existing in the cytoplsm nd some nucler genes [30]. The chrcteristics of the toxin identify five types of killer ctivity mong the strins of this species: K1, K2, K3, KT28, nd K3 [28]. Enologiclly, K2 is the most importnt, becuse its production nd stbility fll within the ph vlues of 2.9 to 4.9, typicl of the musts nd wines; it is the type most commonly found in the yests isolted [23]. The incidence of the killer fctor in spontneous fermenttions hs been studied in most wine regions [7,10,14,15,17,31]. There re two importnt considertions for the wine industry. The first is the competition between killer nd sensitive yests, which hs been relted to the greter implnttion of the former nd even to slowing down of the fermenttion [16,26,32]. Numerous studies im to estblish series of killer/sensitive reltions or rtios, from which the effect of the killer toxin would 1 Deprtmento de Agricultur, Universidd de L Rioj, Avd. de l Pz, 105, Logroño, L Rioj, Spin; 2 Centro de Investigción y Desrrollo Agrrio de L Rioj, P.O. Box 433, Logroño, L Rioj, Spin. *Corresponding uthor [Emil: n-ros.gutierrez@d.unirioj.es] Mnuscript submitted September 1999; revised October 2000, April 2001 Copyright 2001 by the Americn Society for Enology nd Viticulture. All rights reserved. be detectble nd which could give rise to different fermenttion problems. However, there is lck of greement mong these studies. Herd nd Fleet [14] detected the killer effect from K:S proportions of 1:1 nd Brre [2] t proportions of 1:50. Other uthors gve proposed rtios of 25:1 nd 100:1 [33]. The second considertion is the widespred prctice of inocultion of winery tnks with selected yests. The killer phenotype of these strins hs been considered s prcticlly indispensble criterion to ensure the success of the inocultion [33], especilly in those regions where the killer phenotype is found mong the ntive yests flor [15,20]. Some enologists re concerned bout the potentil inhibitory role of killer fctor, prticulrly in res were high proportion of vinifictions occur vi spontneous fermenttion. This concern is bsed on the modifiction cused to winery ecosystems if killer strins re used [8]; the sensitive utochthonous flor my be eliminted, thus preventing their prticiption in other vinifictions. This study investigtes the importnce of the killer phenotype in spontneous fermenttions within the L Rioj DOC, Spin. Smples were tken over two consecutive yers from 11 wineries tht hve never inoculted with commercil yests. To investigte the prevlence of killer yests in winery ecosystem, severl spontneous fermenttions were lso monitored over five consecutive yers t the Centre for Experimenttion nd Agriculturl Development (CIDA), built in 1994 in L Rioj. Unlike the other wineries studied, CIDA does use inocultion. Mterils nd Methods Vinifictions. Commercil wineries in the Rioj DOC: Smples were tken from 11 wineries, locted in the vrious 352

2 Killer Yests 353 Tble 1 Loction nd vinifiction process used t the different wineries studied. Subzone Loction Ref. Vinifiction type Rioj Bj Aldenuev de Ebro 1 Destemming nd crushing Tudelill 2 Crbonic mcertion El Villr de Arnedo 3 Destemming nd crushing Murillo de Río Lez 4 Destemming nd crushing Logroño 5 Destemming nd crushing Rioj Alt Bdrán 6 Destemming nd crushing Alesón 7 Crbonic mcertion Briones 8 Crbonic mcertion Hro 9 Destemming nd crushing Rioj Smniego 10 Crbonic mcertion Alves Villbuen 11 Crbonic mcertion Reference number corresponds to winery number in study. subzones of the Rioj DOC: the Rioj Alt, Rioj Bj, nd Rioj Alves (Tble 1). The smples were collected for two consecutive yers, except in the cse of wineries 3 nd 4, which were only included in the 1998 study. The vinifictions were crried out using the spontneous method, strting from mixtures of severl uthorized grpe vrieties, with Temprnillo being predominnt in ll cses. In ech cse, one tnk ws chosen for smple collection. The vinifictions were crried out by the destemming nd crushing method or by crbonic mcertion (whole grpe), depending on the winery involved (Tble 1). None of the wineries hd ever used commercil yests. No significnt differences in the hygiene conditions existed mong the wineries. The CIDA experimentl winery: The vinifictions were crried out from white must of the Viur vriety over five consecutive yers, from 1994 to The fermenttions took plce in 50-L stinless steel tnks t temperture of 18 C. The number of tnks studied vried ech yer: 12 in 1994, 10 in 1995, 6 in 1996, nd 7 in 1997 nd The fermenttions studied were spontneous, but inocultion ws crried out on other tnks t CIDA with different commercil yests, which hd killer or neutrl phenotype. During this study (1994 to 1998), only six types of commercil yests were used. The mtdna restriction profiles of these strins were known. Isoltion of the yests. At the CIDA winery, smples were tken from the tnks t two stges, the vigorous fermenttion (when the must density ws pproximtely 1.025) nd its end, prt from 1994, when smpling ws crried out only t the vigorous stge. Three smples were tken from the other wineries t the following stges: 24 hr, vigorous fermenttion (density 1.025), nd end of fermenttion. The smples were identified s 1 to 11, s shown in Tble 1, which lso indictes the vinifiction procedure used in ech cse. All smples were collected in sterile bottles nd tken to the lbortory for processing. Seril dilutions were mde nd the smples were seeded onto pltes contining chlormphenicol glucose gr medium (Biokr Dignostic, Beuvis, Frnce). Pltes contining between 30 nd 300 colonies were exmined; 10 colonies from ech smple were rndomly selected. There were 720 isoltes from the CIDA winery nd 600 isoltes from the commercil wineries (270 in 1997 nd 330 in 1998). Determintion of killer phenotype in the isoltes. The killer phenomen of the 1320 isoltes were determined using pltes contining YEPD medium (1% yest extrct, 2% peptone, 2% glucose, nd 2% gr) buffered t ph 4.2 with citric cid-phosphte buffer (21% citric cid nd 28% KH 2 PO 4 ). All isoltes were trnsferred to pltes seeded with reference strin of sensitive S. cerevisie (Spnish Cultivted Type Collection, ref. 1890), nd incubted t 22 C for two dys. The tested strin ws designted killer yest if the streked strin ws surrounded by cler zone of inhibition. To test the sensitive nd neutrl phenotypes, reference strin of killer S. cerevisie (Spnish Cultivted Type Collection, ref. 1893) ws trnsferred to pltes seeded with ech of the nonkiller isoltes. If the killer strin ws surrounded by cler zone of inhibition, then the strin ws designted s sensitive. If there ws no cler zone of inhibition, then the strin ws designted s neutrl. Identifiction of the isoltes. Individul strins were identified by mitochondril DNA (mtdna) restriction nlysis. Yest cells were grown overnight in culture of 5 ml YEPD. DNA extrction nd mtdna restriction ptterns were determined by the methodology described by Querol et l. [21]. The DNA of the 1320 colonies ws digested with the restriction endonuclese Alu I, ccording to the instructions of the supplier (Boehringer Mnnheim, Germny). Some strins were lso digested with the endonucleses Hinf I nd Rs I. The restriction frgments were seprted by electrophoresis (Bio-Rd, Hercules, CA; DNA Sub Cell) in 1% grose gels in 0.5x TBE buffer (45 mm Trisborte, 1mM EDTA, ph8), nd visulized in UV trnsillumintor fter ethidium bromide (5µg/mL) stining for 10 min. Results nd Discussion Identifiction of the isoltes. The mitochondril DNA restriction nlysis of the isoltes from the commercil wineries showed tht 24% of the strins hd profile belonging to non- S. cerevisie yests. Ninety percent of these strins cme from the smples tken fter 24 hr. However, no non-s. cerevisie yests were found in the CIDA smples (except for the 1997 hrvest), possibly due to not hving tken smples t 24 hr. The 1997 hrvest ws problemtic in mny Spnish wine regions, including the Rioj DOC. Two non-s. cerevisie clones represented 50% of ll the isoltes found for the vigorous fermenttion phse in 1997, which my be directly relted to the problems suffered. Among the isoltes belonging to the S. cerevisie genus, mtdna restriction nlysis reveled 73 different profiles, which cn be interpreted s different clones [13], for the five yers of the CIDA. The commercil wineries gve totl of 205 different ptterns (89 in 1997 nd 116 in 1998). It is remrkble tht only three out of six inoculted strins ppered in the tnks of the CIDA winery, but in very low percentge (dt not shown). Seventy spontneous yest profiles ppered, despite six strins of yests used to inoculte in other tnks. Therefore, concerns bout the modifiction in the winery ecosystem fter inocul-

3 354 Gutiérrez et l. tion with killer yests should not prevent inocultions when needed. Study of the killer chrcter. Among the clones clssified s non-s. cerevisie, 72.5% showed the neutrl phenotype nd 27.5% the sensitive phenotype. No killer phenotypes were found, greeing with the results of other uthors [3,22], despite the fct tht the existence of the killer phenotype hs been demonstrted in other enologicl yests prt from S. cerevisie [1,15]. Tble 2 shows the distribution mong the S. cerevisie isoltes of the different killer phenotypes by winery nd DOC subzone. The dt is the verge of the three smples obtined in ech winery. The distribution of the killer fctor is linked to the zone nd techniques used: killer strins ppered only in the wineries of the Rioj Bj. In the Rioj Alt nd Rioj Alves, the sensitive phenotype ws clerly dominnt, to the point of representing 100% in some yers nd wineries. There were three wineries (4, 5, nd 6) whose dominnt phenotype in some yers ws the neutrl; two of these wineries re locted on the border between the two subzones of the Rioj Alt nd Bj. This dt grees with study on the killer fctor in the Rioj DOC conducted in 1993, which investigted six wineries in the re [27]; the killer phenotype ws found only in the two wineries locted in the Rioj Bj. In regrd to elbortion technique, those wineries using the trditionl crbonic mcertion system showed cler dominnce of the sensitive phenotype, even in the Tudelill winery (number 2) in the Rioj Bj. The wineries using destemming showed greter vrition: different dominnt phenotypes ppered depending on the yer nd the winery. The results show distinction in the presence of killer yest ssocited with the two vinifiction systems. Severl possible explntions could exist: the crbonic mcertion process by some unknown mechnism prevented the killing ction from being effective (inctivting the killer toxin), so tht the killer yests hd no competitive dvntge. Alterntively, the specific conditions required for the crbonic mcertion process (nerobiosis, high temperture nd ph, presence of stems) my inctivte the K2 toxin nd my influence the proportion of killer cells present during fermenttion, s demonstrted for other fctors [1,18,22]. These results were shown by our reserch group in study of the 1994 nd 1995 hrvests, in which the influence of the must clrifiction ws clerly seen on this distribution [12]. Another possible explntion could be tht the crbonic mcertion process ctively selects ginst the killer yests, ssuming tht killer yests were in fct present in ll the wineries concerned. Bsed on the current literture, it is difficult to understnd how the process of crbonic mcertion could ctively remove killer yest from tht environment. This possibility could be tested by dding killer yests into the crbonic mcertion process. The distribution of killer, neutrl, nd sensitive strins in S. cerevisie t the CIDA winery over five consecutive yers nd t different stges of the fermenttions is shown in Tble 3. In generl, coexistence of the three phenotypes ws found t the sme fermenttive stges; no consistent trends were seen for ny phenotype s the fermenttions progressed. Where multiple smples were tken, the killer strins showed slight increse for ll yers other thn The proportion of neutrl strins incresed t the end of fermenttion in 1995, but fell in 1996 nd In 1996, 1997, nd 1998, the sensitive strins incresed with fermenttion progress. Other studies [20] hve not found coexistence between killer nd sensitive strins t the sme fermenttive stge nd hve indicted tht the sensitive strins re eliminted by the killer toxin; this is the sme reson given by Cortz [9] to explin the increse found in the killer strins t the end of the fermenttion. However, Cnsdo [6] reported reduction in the killer frequency, explined s the degrdtion or inctivtion of the toxins t the end of the fermenttion, thus llowing greter survivl rte of the sensitive strins. Our results do not show one cler tendency nd vry with the yer studied. The killer strins do not hve mjor dvntge in their implnttion, s Tble 2 Frequency (%) of the killer, neutrl, nd sensitive strins found in the S. cerevisie isoltes from wineries of DOC Rioj; 1997 nd Averge dt for the whole fermenttion (24 hr, vigorous, nd end of fermenttion). Rioj Bj Rioj Alt Rioj Alves 1 2 b b 8 b 9 10 b 11 b 1997 Killer 63 c - * * Neutrl 37 3 * * Sensitive - 97 * * Killer Neutrl Sensitive Wineries: Aldenuev de Ebro (1); Tudelill (2); El Villr de Arnedo (3); Murillo (4); Logroño (5); Bdrán (6); Alesón (7); Briones (8); Hro (9); Smniego (10); Villbuen (11). b Wineries with production by crbonic mcertion. c Bold numbers correspond to the dominnt phenotype ppering in every winery. *Included in the study for 1998, no dt for Tble 3 Frequency (%) of the different S. cerevisie phenotypes t the CIDA winery over five-yer period from 1994 to Dt given is the verge from ll the tnks tested ech yer (12 in 1994, 10 in 1995, 6 in 1996, nd 7 in 1997 nd 1998) VF EF AF VF EF AF VF EF AF VF EF AF VF EF AF Killer 64 nd b Neutrl 36 nd Sensitive 0 nd VF: vigorous fermenttion; EF: end fermenttion; AF: verge fermenttion. b nd: no dt

4 Killer Yests 355 they did not clerly impose themselves in ny cse (except for 1994, where dt ws tken only for the vigorous fermenttion) (Tble 3). Furthermore, the sensitive strins were ble to develop, in some cses (1998) even becoming dominnt. These results do not gree with other studies [3,20], which indicte tht n inoculum of killer strins t 5% or greter, growing with sensitive strins, is cpble of dominting nd even eliminting the sensitive popultion over the course of the fermenttion. Other uthors [33] indicte tht much higher level of killer strin is required to ensure tht their effect be noticeble. These differences in the results my be due to the presence of the neutrl strins, which, ccording to Silv [11], my protect the sensitive strins ginst the killer toxin. The vrition in phenotype distribution with respect to the killer chrcter does not necessrily imply tht the S. cerevisie strins tht drive the fermenttion re replced. This study found geneticlly identicl strins (identicl mtdna restriction profiles derived with the Alu I enzyme) with differing killer phenotypes. These strins were subjected to further mtdna restriction nlysis using the Hinf I (Figure 1) nd Rs I (dt not shown) endonucleses; both cses showed tht they re the sme clones, indicting tht geneticlly equivlent strins cn exist, with differing killer phenotypes. The ctivity of the killer toxin, nd resistnce to it, depend on severl fctors such s the ph, temperture, growth conditions, nd the medium [5,18,34]. These fcts could explin why the sme strins cn present different phenotypes over the fermenttion process, s the fctors will chnge s it progresses. It hs lso been noted tht the killer phenotype is not stble in yests nd cn be lost during conservtion over long periods. Tiourin [29] proposed long periods, even up to 25 yers; in our lbortory, we hve seen tht this effect my tke plce in much shorter time (unpublished results) Figure 1 Mitochondril DNA electrophoresis of some S. cerevisie strins isolted from the CIDA winery, using the restriction enzyme Hinf I. Lne 1 corresponds to phge lmbd cut with EcoRI nd Hind III. Lnes 2, 3, 4, 6, nd 7, with identicl restriction profile but with different killer phenotype, correspond to strin XVII from the 1995 vintge. Lne 5 corresponds to strin III coming from the sme yer. Conclusions The dt reported here question the importnce tht hs been given to the killer phenotype, minly in the process of strin selection for lter commerciliztion. We hve shown tht killer strins llow development of sensitive strins nd tht the lter my even dominte, perhps due to their better dpttion to the vinifiction conditions. As result, we gree with D Silv [11] tht from the enologicl point of view, yests should be selected on the bsis of their enologicl chrcteristics nd independently of their killer ctivity. A killer phenotype might not confer n dvntge in vinifiction, depending on multiple fctors. Sensitive strins re cpble not only of mintining but lso of incresing their proportion, to rech very high percentges close to 100%, s found in the Rioj Alt nd the Rioj Alves. This occurs yer fter yer, following nturl selection over time for ech winery nd re, nd suggests tht these strins hve some selective dvntge over the killer phenotypes, bsed on their dpttion to the vinifiction conditions (the must chrcteristics nd technique used). Regrding the concern tht seeding with commercil killer strins will lter the sensitive flor of wineries, the results obtined from the CIDA winery (Tble 3) should be emphsized: the dominnt phenotype with respect to the killer fctor of the predominnt strins vried from one yer to the next. The killer ws dominnt in 1994, the neutrl in 1995, 1996, nd 1997, nd the sensitive in 1998, even in the presence of killer strins. Despite the fermenttions studied being spontneous, other works conducted t CIDA hd involved inocultion with commercil killer yests. However, sensitive strins were ble to develop nd even dominte in some spontneous fermenttions. Literture Cited 1. Angulo, L., A.M. Sborido, C. Lem, J. Freire, J.E. López, nd J.A. Vicente. Selección y diferencición enológic de ceps de Scchromyces cerevisie islds durnte l vinificción en l comrc de O Bixo Miño (Suroeste de Glici). Vitic. Enol. Profesionl 27:28-36 (1993). 2. Brre, P. Le fcteur killer. Les cquisitions récentes en microbiologie du vin. Incidences sur les propiétés et les ltértions du vin. B. Doneche (Ed.), pp Tec & Doc Lvoisier, Pris (1992). 3. Brre, P. Rôle du fcteur killer dns l concurrence entre souches de levures. Bull. OIV 53: (1980). 4. Bevn, E.A., nd M. Mkower. The physiologicl bsis for the killer chrcter in yest. In Genetics Tody, Eleventh Interntionl Congress on Genetics. Vol 1. S.J. Gerts (Ed.), pp Pergmon Press, Oxford (1963). 5. Bussey, H. Effects of yest killer fctor on sensitive cells. Nture New Biol. 235:73-75 (1972). 6. Cnsdo, J., E. Longo, D. Agrelo, nd T.G. Vill. Levdurs socids procesos de fermentción espontáne de vinos de Ribeiro. Análisis del homo/heterotlismo y sistem killer de ls ceps de Scchromyces cerevisie. Microbiol. SEM. 5:79-88 (1989). 7. Ciolfi, G. Distribuzione dei fenotipi killer, neutro, sensibile nel corso di fermentzioni spontnee. Riv. Vitic. Enol. 4: (1985). 8. Constntí, M., M. Poblet, I. Arol, A. Ms, nd J.M. Guillmón. Anlysis of yest popultions during lcoholic fermenttion in newly estblished winery. Am. J. Enol. Vitic. 48(3): (1997).

5 356 Gutiérrez et l. 9. Cortz, G., R.A. Musmnno, S. Cresti, P. Vgnoli, nd T. di Mggio. L evoluzione dell popolzione di lieviti durnte l fermentzione. Vignevini 9:24-27 (1992). 10. Cuinier, C., nd C. Gros. Enquête sur l réprtition des levures killer en Frnce. Vigne Vine 318:25-27 (1983). 11. D Silv, G.A. The occurrence of killer, sensitive, nd neutrl yests in Brzilin Riesling Itlico grpe must nd the effect of neutrl strins on killing behviour. Appl. Microbiol. Biotechnol. 46: (1996). 12. Epifnio, S.I., A.R. Gutiérrez, M.P. Sntmrí, nd R.López. The influence of enologicl prctices on the selection of wild yest strins in spontneous fermenttion. Am. J. Enol. Vitic. 50(2): (1999). 13. Guillmón, J.M. Estudio ecológico de l fermentción lcohólic medinte l utilizción de mrcdores moleculres. Thesis, Universidd de Vlenci (1996). 14. Herd, G.M., nd G.H. Fleet. Occurrence nd growth of killer yests during wine fermenttions. Appl. Environ. Microbiol. 51: (1987). 15. Hidlgo, P., nd M. Flores. Occurrence of the killer chrcter in yest ssocited with Spnish wine production. Food Microbiol. 11: (1994). 16. Jcobs, C.J., nd H. J. J. Vn Vuuren. Effects of different killer yests on wine fermenttion. Am. J. Enol. Vitic. 42: (1991). 17. Kitno, K., M. Sto, T. Shimzki, nd S. Hr. Occurrence of wild killer yest in Jpnese wineries nd their chrcteristics. J.Ferment. Tech. 62:1-6 (1984). 18. Medin, K., F. Crru, O. Giol, nd N. Brcesco. Nitrogen vilbility of grpe juice limits killer yest growth nd fermenttion ctivity during mixed-culture fermenttion with sensitive commercil yest strins. Appl. Environ. Microbiol. 63(7): (1997). 19. Nguyen, H., nd G. Pnon. The yest Metschnikowi pulcherrim hs n inhibitory effect ginst vrious yest species. Sci. Aliments 18: (1998). 20. Perez, F., J.A. Mteos, M.E. Vldés, nd M. Rmírez. Estudio del fctor killer en ceps de levdurs islds de vinos extremeños privte. Third Symposium on Vine-growing in the Alentejo, Evor, pp ATEVA, Portugl (1995). 21. Querol, A., E. Brrio, T. Huert, nd D. Rmón. Moleculr monitoring of wine fermenttions conduced by dry yests strins. Appl. Environ. Microbiol. 58: (1992). 22. Quesd, M.P., nd A. Mrtínez. Respuest enológic de ceps de levdurs vínics Scchromyces cerevisie de l región de Murci con fenotipos killer, neutro y sensible. Third Symposium on Vine-growing in the Alentejo, Evor, pp ATEVA, Portugl (1995). 23. Quesd, M.P., A. Mrtínez, J.M. López, nd F. Mrín. Influenci de l toxin killer durnte l vinificción. Vitivinicultur 5/6:52-56 (1995). 24. Rdler, F., nd C. Knoll. Formtion of killer toxin by piculte yest nd interference with fermenttion. Vitis 27: (1988). 25. Rdler, F. Les fcteurs killer des levures. Bull. OIV 53: (1980). 26. Rosini, G. Effet d une levure Killer de Scchromyces cerevisie sur une souche de levure sensible de l même espece, non productrice de H 2 S et sélectionée pour l vinifiction dns un milieu de culture mixte. Bull. OIV 58( ): (1985). 27. Senz-Olzbl, M. Fctor killer en levdurs de Rioj. Thesis, Universidd de L Rioj, Spin (1994). 28. Shimizu, K. Killer yest. In Wine Microbiology nd Biotechnology. G.H. Fleet (Ed.), pp Hrwood Acdemic Publishers, Switzerlnd (1993). 29. Tiourin, R.H., N.I. Bourjn, nd T.K. Skrikov. Emploi de cultures pures du phénotype killer dns l fermenttion de moûts de risin. Bull. OIV 53: (1980). 30. Tipper, D.J., nd K.A. Bostin. Double-strnded ribonucleic cid killer systems in yests. Microbiol. Rev. 48: (1984). 31. Vgnoli, P., R.A. Musmnno, S. Cresti, T. di Mggio, nd G. Cortz. Occurrence of killer yests in spontneous wine fermenttions from the Tuscny region of Itly. Appl. Environ. Microbiol. 59: (1993). 32. Vn Vuuren, H.J.J., nd B.D. Wingfield. Killer yests-cuse of stuck fermenttions in wine cellr. S. Afr. J. Enol. Vitic. 7: (1986). 33. Vn Vuuren, H.J.J., nd J.C. Jcobs. Killer yests in the wine industry: A review. Am. J. Enol. Vitic. 43: (1992). 34. Woods, D.R., nd E.A. Bevn. Studies on the nture of killer fctors produced by Scchromyces cerevisie. J. Gen. Microbiol. 51: (1968).