Wine yeast selection in the Iberian Peninsula: Saccharomyces and non-Saccharomyces as drivers of innovation in Spanish and Portuguese wine industries.

Yeast selection for the wine industry in Spain started in 1950 for the understanding of the microbial ecology, and for the selection of optimal strains to improve the performance of alcoholic fermentation and the overall wine quality. This process has been strongly developed over the last 30 years, firstly on Saccharomyces cerevisiae, and, lately, with intense activity on non-Saccharomyces. Several thousand yeast strains have been isolated, identified and tested to select those with better performance and/or specific technological properties. The present review proposes a global survey of this massive ex-situ preservation of eukaryotic microorganisms, a reservoir of biotechnological solutions for the wine sector, overviewing relevant screenings that led to the selection of strains from 12 genera and 22 species of oenological significance. In the first part, the attention goes to the selection programmes related to relevant wine-producing areas (i.e. Douro, Extremadura, Galicia, La Mancha and Uclés, Ribera del Duero, Rioja, Sherry area, and Valencia). In the second part, the focus shifted on specific non-Saccharomyces genera/species selected from different Spanish and Portuguese regions, exploited to enhance particular attributes of the wines. A fil rouge of the dissertation is the design of tailored biotechnological solutions for wines typical of given geographic areas.

Growth and metabolite production of a grape sour rot yeast-bacterium consortium on different carbon sources.

The present study was designed to evaluate possible sugar-based trophic interactions between acetic acid bacteria (AAB) and non-Saccharomyces yeasts (NSY) involved in table grape sour rot, a disease in which berries spoilage is caused by the accumulation of several microbial metabolites. Acetobacter syzygii LMG 21419 (As) and Candida zemplinina CBS 9494 (Cz), a simplified AAB-NSY association responsible for table grape sour rot, grew differently in a minimal medium (YP) supplemented with glucose, ethanol, acetic and gluconic acid under monoculture conditions. In As -Cz co-culture media, after 24 h of incubation, As showed high relative abundance in YP-ethanol, whereas Cz was the dominant strain in YP-glucose medium. Co-culture in YP-glucose showed that glucose was converted into ethanol by Cz that, in turn, promoted the growth of As population. Gluconic acid was the main bacterial metabolite from glucose in monoculture, whereas acetic acid putatively derived from ethanol oxidation was found only in co-culture. However, gluconic acid showed inhibitory effect against As whereas acetic acid mainly inhibited Cz. Negative effects of both metabolites were mitigated in the glucose-supplemented medium. The results suggest a possible metabolic- based temporal succession between AAB and NSY during grape sour rot development. At the begin of sour rot, low glucose concentration promotes NSY producing ethanol, then, the AAB could take advantage from the oxidation of ethanol into acetic acid, becoming the dominant microbial sour rot population during the late stages of the process.

A Response Surface Methodology study on the role of factors affecting growth and volatile phenol production by Brettanomyces bruxellensis ISA 2211 in wine.

The present study was aimed at determining the effect of glucose, ethanol and sulphur dioxide on the growth and volatile phenol production by Brettanomyces bruxellensis in red wines using a response surface methodology approach. Sulphur dioxide proved to have a significant (p < 0.05) negative linear and quadratic effect on growth and 4-ethylphenol production. Concentrations of sulphur dioxide higher than 20 mg L(-1), at pH 3.50, induced immediate loss of cell culturability under growth permissive levels of ethanol. Under high ethanol concentrations (14% v/v), the lag phase increased from 3 to 10 days, growth being fully arrested at 15% (v/v). Glucose up to 10 g L(-1) was found to be a significant factor (quadratic level) in biomass increase under low ethanol (<12.5% v/v) and low sulphite concentrations. However, when cells were inactivated by sulphur dioxide and ethanol, glucose (up to 10 g L(-1)) did not prevent cell death. Production of more than 50 µg L(-1) day(-1) of 4-ethylphenol was only observed in the presence of high numbers (10(6) CFU mL(-1)) of culturable cells, being stimulated by increasing glucose concentrations.

The microbial ecology of wine grape berries.

Grapes have a complex microbial ecology including filamentous fungi, yeasts and bacteria with different physiological characteristics and effects upon wine production. Some species are only found in grapes, such as parasitic fungi and environmental bacteria, while others have the ability to survive and grow in wines, constituting the wine microbial consortium. This consortium covers yeast species, lactic acid bacteria and acetic acid bacteria. The proportion of these microorganisms depends on the grape ripening stage and on the availability of nutrients. Grape berries are susceptible to fungal parasites until véraison after which the microbiota of truly intact berries is similar to that of plant leaves, which is dominated by basidiomycetous yeasts (e.g. Cryptococcus spp., Rhodotorula spp. Sporobolomyces spp.) and the yeast-like fungus Aureobasidium pullulans. The cuticle of visually intact berries may bear microfissures and softens with ripening, increasing nutrient availability and explaining the possible dominance by the oxidative or weakly fermentative ascomycetous populations (e.g. Candida spp., Hanseniaspora spp., Metschnikowia spp., Pichia spp.) approaching harvest time. When grape skin is clearly damaged, the availability of high sugar concentrations on the berry surface favours the increase of ascomycetes with higher fermentative activity like Pichia spp. and Zygoascus hellenicus, including dangerous wine spoilage yeasts (e.g. Zygosaccharomyces spp., Torulaspora spp.), and of acetic acid bacteria (e.g. Gluconobacter spp., Acetobacter spp.). The sugar fermenting species Saccharomyces cerevisiae is rarely found on unblemished berries, being favoured by grape damage. Lactic acid bacteria are minor partners of grape microbiota and while being the typical agent of malolactic fermentation, Oenococcus oeni has been seldom isolated from grapes in the vineyard. Environmental ubiquitous bacteria of the genus Enterobacter spp., Enterococcus spp., Bacillus spp., Burkholderia spp., Serratia spp., Staphylococcus spp., among others, have been isolated from grapes but do not have the ability to grow in wines. Saprophytic moulds, like Botrytis cinerea, causing grey rot, or Aspergillus spp., possibly producing ochratoxin, are only active in the vineyard, although their metabolites may affect wine quality during grape processing. The impact of damaged grapes in yeast ecology has been underestimated mostly because of inaccurate grape sampling. Injured berries hidden in apparently sound bunches explain the recovery of a higher number of species when whole bunches are picked. Grape health status is the main factor affecting the microbial ecology of grapes, increasing both microbial numbers and species diversity. Therefore, the influence of abiotic (e.g. climate, rain, hail), biotic (e.g. insects, birds, phytopathogenic and saprophytic moulds) and viticultural (e.g. fungicides) factors is dependent on their primary damaging effect.

Evaluation of the inhibitory effect of dimethyl dicarbonate (DMDC) against wine microorganisms.

Several microbial species associated with wine were challenged against increasing concentrations of dimethyl dicarbonate (DMDC). The concentration inducing complete cell death upon addition to red wine was regarded as the minimum inhibitory concentration (MIC). In dry red wines with 12% (v/v) ethanol and pH 3.50, the inactivation depended on the initial cell concentration. For an initial inoculum of 500 CFU/ml, the MIC of the yeasts species Schizosaccharomyces pombe, Dekkera bruxellensis, Saccharomyces cerevisiae and Pichia guilliermondii was 100mg/l. The most sensitive strains belong to Zygosaccharomyces bailii, Zygoascus hellenicus and Lachancea thermotolerans, with MIC of 25mg/l DMDC. For inoculation rates of about 10(6)CFU/ml, the maximum dose of DMDC legally authorized (200mg/l) was not effective against the most resistant species. The addition of 100mg/l potassium metabisulphite (PMB), equivalent to 1mg/l molecular sulphur dioxide, increased the inactivation effect of 100mg/l DMDC over initial yeast populations of 10(6)CFU/ml but did not fully kill S. pombe and S. cerevisiae. Lactic acid and acetic acid bacteria were not killed by the addition of 300 mg/l of DMDC. Trials performed in wines before bottling showed that in most samples indigenous bacterial populations were not affected by 200mg/l DMDC. Therefore, under winery practice, DMDC at the maximum dose legally permitted may be regarded as an efficient preservative to control low contamination rates of yeasts but ineffective against lactic acid and acetic acid bacteria.

Survival patterns of Dekkera bruxellensis in wines and inhibitory effect of sulphur dioxide.

The wine spoilage yeast species Dekkera bruxellensis, after inoculation in red wines, displayed three survival patterns characterized by: i) initial lag phase followed by growth and sequential death; ii) initial death phase leading to reduced viable counts followed by growth and sequential death; and iii) death phase leading to complete loss of viability. These survival patterns were observed for the same strain in different dry red wine blends with 12% (v/v) ethanol and pH 3.50, in the absence of free sulphur dioxide. For the same wine blend, these patterns also varied with the tested strain. Under laboratory conditions the addition of 150 mg/l of potassium metabisulphite (PMB) to dry red wine with 12% (v/v) ethanol and pH 3.50 reduced initial cell counts by more than 6 logarithmic cycles, inducing full death within less than 24 h. Winery trials showed that D. bruxellensis blooms were only prevented in the presence of about 40 mg/l of free sulphur dioxide in dry red wine, with 13.8% (v/v) ethanol and pH 3.42, matured in oak barrels. These different amounts of PMB and sulphur dioxide corresponded to about 1 mg/l of molecular sulphur dioxide. Our results therefore demonstrate that the control of populations of D. bruxellensis growing in red wine can only be achieved under the presence of relatively high doses of molecular sulphur dioxide.

Ascomycetous yeast species recovered from grapes damaged by honeydew and sour rot.

AIMS: To identify ascomycetous yeasts recovered from sound and damaged grapes by the presence of honeydew or sour rot. METHODS AND RESULTS: In sound grapes, the mean yeast counts ranged from 3.20 +/- 1.04 log CFU g(-1) to 5.87 +/- 0.64 log CFU g(-1). In honeydew grapes, the mean counts ranged from 3.88 +/- 0.80 log CFU g(-1) to 6.64 +/- 0.77 log CFU g(-1). In sour rot grapes counts varied between 6.34 +/- 1.03 and 7.68 +/- 0.38 logCFU g(-1). Hanseniaspora uvarum was the most frequent species from sound samples. In both types of damage, the most frequent species were Candida vanderwaltii, H. uvarum and Zygoascus hellenicus. The latter species was recovered in high frequency because of the utilization of the selective medium DBDM (Dekkera/Brettanomyces differential medium). The scarce isolation frequency of the wine spoilage species Zygosaccharomyces bailii (in sour rotten grapes) and Zygosaccharomyces bisporus (in honeydew affected grapes) could only be demonstrated by the use of the selective medium ZDM (Zygosaccharomyces differential medium). CONCLUSIONS: The isolation of several species only from damaged grapes indicates that damage constituted the main factor determining yeast diversity. The utilization of selective media is required for eliciting the recovery of potentially wine spoilage species. SIGNIFICANCE AND IMPACT OF THE STUDY: The impact of damaged grapes in the yeast ecology of grapes has been underestimated.

An inter-laboratory evaluation of selective media for the detection and enumeration of yeasts from blue-veined cheese.

Five countries representative of laboratories 1-5 evaluated 11 different selective media, designed to suppress mould and bacterial growth and support yeasts growth, for the recovery of yeast populations from blue veined cheeses. In addition, qualitative results were also incorporated. The yeast enumeration values were subjected to statistical analysis using analysis of variance (ANOVA) and the Tukey-Kramer multiple comparison test. With the exception of Laboratory 3, none of the other laboratories was successful in recovering yeasts on all the media. Six of the media proved inadequate for the enumeration of yeasts in the mould invested environment and were therefore omitted from statistical analysis. No significant differences in quantitative data obtained on Rose-Bengal Chloramphenicol Agar (RBCA), Dichloran Rose-Bengal Chloramphenicol Agar (DRBC), Dichloran 18% Glycerol Agar (DG18), and Malt extract agar supplemented with NaCl and oxytetracycline (MES) were detected by four of the collaborating laboratories whereas one laboratory found RBCA to be superior for yeast enumeration. DG18 and Malt Extract Agar with Biphenyl (MEB), however, were ranked superior based on qualitative results compared to the other media, attributed to distinctive individual yeast colonies and mould inhibition. RBCA, DRBC, DG18, and MES on the other hand, all proved to be adequate in supporting yeast colony development for quantitative analysis in samples obtained from blue veined cheeses.

Spoilage yeasts in the wine industry.

Yeasts play a central role in the spoilage of foods and beverages, mainly those with high acidity and reduced water activity (a(w)). A few species are capable of spoiling foods produced according to good manufacturing practices (GMPs). These can survive and grow under stress conditions where other microorganisms are not competitive. However, many of the aspects determining yeast spoilage have yet to be clarified. This critical review uses the wine industry as a case study where serious microbiological problems are caused by yeasts. First, the limitations of the available tools to assess the presence of spoilage yeasts in foods are discussed. Next, yeasts and factors promoting their colonisation in grapes and wines are discussed from the ecological perspective, demonstrating that a deeper knowledge of vineyard and winery ecosystems is essential to establish the origin of wine spoilage yeasts, their routes of contamination, critical points of yeast infection, and of course, their control. Further, zymological indicators are discussed as important tools to assess the microbiological quality of wines, although they are rarely used by the wine industry. The concepts of the susceptibility of wine to spoilage yeasts and wine stability are addressed based on scientific knowledge and industrial practices for monitoring yeast contamination. A discussion on acceptable levels of yeasts and microbiological criteria in the wine industry is supported by data obtained from wineries, wholesalers, and the scientific literature.Finally, future directions for applied research are proposed, involving collaboration between scientists and industry to improve the quality of wine and methods for monitoring the presence of yeast.

Development and use of a new medium to detect yeasts of the genera Dekkera/Brettanomyces.

AIMS: The objectives of this work were to develop a selective and/or differential medium able to efficiently recover Dekkera/Brettanomyces sp. from wine-related environments and to determine the relationship between these yeasts and the 4-ethylphenol content in a wide range of wines. METHODS AND RESULTS: The selectivity of the developed medium was provided by the addition of ethanol, as single carbon source, and cycloheximide. The inclusion of bromocresol green evidenced acid-producing strains. The inclusion of p-coumaric acid, substrate for the production of 4-ethylphenol, enabled the differentiation by smell of Dekkera/Brettanomyces sp. from all other yeast species growing in the medium. The medium was used either by plating after membrane filtration or by the Most Probable Number (MPN) technique. In 29 white and 88 red randomly collected wines, these yeasts were found only in red wines at levels up to 2500 MPN ml-1, but constituted less than 1% of the total microbial flora. In red wines, 84% showed detectable amounts of 4-ethylphenol up to 4430 microg l-1 while 28% of the white wines showed detectable levels up to 403 microg l-1. CONCLUSION: The use of the medium proposed in this work evidenced the presence of low relative populations of Dekkera/Brettanomyces sp. even in wines contaminated by fast-growing yeasts and moulds. SIGNIFICANCE AND IMPACT OF THE STUDY: Further ecological studies on Dekkera/Brettanomyces sp. should take into account the use of highly specific culture media in order to establish their true occurrence in nature.

Characterisation of yeast flora isolated from an artisanal Portuguese ewes' cheese.

The evolution of the yeast flora was studied for an artisanal semi-hard ewes' cheese made from raw milk. Mean log10 yeast counts per gram of cheese body ranged from 2.7 to 6.4, with the higher counts observed after a ripening period of 30 days. The yeast population decreased thereafter and, at the end of curing process, reached values similar to those of the beginning. A total of 344 yeasts strains were randomly isolated from the curd and cheese body during the 60 days long ripening period. Esterase activity was common to almost all isolates (98%) while proteolysis was observed in 12% of the total yeast population. The proportion of strains with positive glucose fermentation increased from 21% in the curd to 75% at the end of the ripening period. A total of 150 isolates representative of the physiological characteristics tested were examined with the API ID 32C system showing different degrees of quality of identification. Only 15% of the strains (23 isolates) were excellently identified being assigned to the species Candida zeylanoides. The most frequent species appeared to be Debaryomyces hansenii (anamorph Candida famata) and Candida intermedia. These two species amounted to 9% of the yeasts in the curd increasing to 86% at the end of the ripening period.

Characterization of yeasts associated with Portuguese pork-based products.

Yeasts from pork-based products, such as country-cured hams and several types of bacon, were studied. Contamination levels ranged from 10(3)-10(9) CFU/g of fat. Identification using a simplified method (SIM) showed that the predominant species were Debaryomyces hansenii, Cryptococcus laurentii, Cryptococcus humicolus, Debaryomyces polymorphus and Pichia guilliermondii. Lipolytic activity was observed in 24 out of 29 of the isolated strains and was not a determinant for attaining high levels of contamination. The analysis of yeast long-chain fatty acid composition for most isolated strains showed a predominance of the acid C18:1 followed by decreasing proportions of the acids C16:0, C18:2, C18:3, C18:0 and C16:1. The fatty acid profiles were roughly similar for all yeasts analyzed suggesting that in these types of pork-based products the yeast lipid composition may have an ecological significance.

Fatty acid profiling: a feasible typing system to trace yeast contamination in wine bottling plants.

The long-chain fatty acid composition of yeast strains was determined for several species associated with the wine industry. The Saccharomyces cerevisiae, Zygosaccharomyces bailii, Saccharomycodes ludwigii, Schizosaccharomyces pombe, Brettanomyces/Dekkera spp., Pichia anomala, Pichia membranaefaciens and Lodderomyces elongisporus species presented distinct fatty acid profiles after multivariate statistical analysis. The Zygosaccharomyces rouxii species showed profiles similar to Zygosaccharomyces bailii. The use of fatty acid profiling in wine bottling plants and wines makes it possible to trace the origin of the strains responsible for spoiling the final product. In one case the origin was found at the outlet of the finishing filter and identified as Zygosaccharomyces bailii. In the other case the source of contamination was discovered in the heads of the filling machine and assigned to the Pichia membranaefaciens species. The results point out the discriminating power and the industrial applicability of the technique described in this work to analyse yeast long-chain fatty acid compositions.