- Top of page
- 2Materials and methods
One thousand six hundred and twenty yeast isolates were obtained from 54 spontaneous fermentations performed from grapes collected in 18 sampling sites of three vineyards (Vinho Verde Wine Region in northwest Portugal) during the 2001–2003 harvest seasons. All isolates were analyzed by mitochondrial DNA restriction fragment length polymorphism (mtDNA RFLP) and a pattern profile was verified for each isolate, resulting in a total of 297 different profiles, that all belonged to the species Saccharomyces cerevisiae. The strains corresponding to seventeen profiles showed a wider temporal and geographical distribution, being characterized by a generalized pattern of sporadic presence, absence and reappearance. One strain (ACP10) showed a more regional distribution with a perennial behavior. In different fermentations ACP10 was either dominant or not, showing that the final outcome of fermentation was dependent on the specific composition of the yeast community in the must. Few of the grape samples collected before harvest initiated a spontaneous fermentation, compared to the samples collected after harvest, in a time frame of about 2 weeks. The associated strains were also much more diversified: 267 patterns among 1260 isolates compared to 30 patterns among 360 isolates in the post- and pre-harvest samples, respectively. Fermenting yeast populations have never been characterized before in this region and the present work reports the presence of commercial yeast strains used by the wineries. The present study aims at the development of strategies for the preservation of biodiversity and genetic resources as a basis for further strain development.
- Top of page
- 2Materials and methods
Traditionally, wine fermentation is carried out in a spontaneous way by indigenous yeast either present on the grapes when harvested or introduced from the equipment and cellar during the vinification process. All recent research agrees that the predominant species on healthy grapes are apiculate yeasts like Hanseniaspora uvarum (and its anamorph form Kloeckera apiculata) and oxidative species such as Candida, Pichia, Kluyveromyces and Rhodotorula. Contrarily, fermentative species of the genus Saccharomyces, predominantly Saccharomyces cerevisiae, occur in extremely low number on healthy undamaged berries or in soils [2–4], while damaged grapes are believed to be an important source of S. cerevisiae. The prevalence of strains belonging to this species is well documented among the wineries resident flora [6–10]. The grape's yeast flora depends on a large variety of factors such as climatic conditions including temperature and rainfalls, geographic localization of the vineyard [4,9], antifungal applications, grape variety and the vineyard's age [12–14], as well as the soil type. Several ecological surveys, using molecular methods of identification, report a large diversity of genetic patterns among the enological fermentative flora. S. cerevisiae strains seem to be widely distributed in a given viticultural region [16–19], can be found in consecutive years [20,21] and there are also strains predominant in the fermenting flora [2,22], which support the notion that specific native strains can be associated with a terroir.
Selected yeast starters are nowadays widely used since they possess very good fermentative and oenological capabilities, contributing to both standardization of fermentation process and wine quality. In the years following the publication of the S. cerevisiae genome sequence, enough evidence was provided showing substantial genetic differences among wine yeast strains [24–26]. Therefore, exploring the biodiversity of indigenous fermentative strains can be an important contribution towards the understanding and selection of strains with specific phenotypes.
The genetic diversity of S. cerevisiae strains has been analyzed by several methods such as karyotyping by pulse field gel electrophoresis, mitochondrial DNA restriction analysis (mtDNA RFLP) [28–31], fingerprinting based on repetitive delta sequences [32,33] and microsatellite genotyping [34–36]. Schuller et al. have recently shown that microsatellite typing, using 6 different loci, an optimized interdelta sequence analysis and RFLP of mitochondrial DNA generated by the enzyme Hinf I had the same discriminatory power. In the present work mtDNA RFLP analysis using Hinf I was applied as genetic marker for the distinction of S. cerevisiae strains.
The aim of the present work was to assess the biodiversity of the fermenting flora found in vineyards belonging to the Vinho Verde Region in order to define strategies for future wine strain selection programs. Another goal was the establishment of a strain collection contributing to the preservation of S. cerevisiae genetic resources.
- Top of page
- 2Materials and methods
In the present work, three vineyards, situated in the Vinho Verde Wine Region, in northwest Portugal, were sampled during the 2001–2003 harvest seasons (Fig. 1). In order to obtain a more detailed picture of fermenting yeast temporal distribution, two sampling campaigns were performed, one before and another after the harvest, in a time frame of about two weeks. A total of 108 grape samples have been planned (six sampling points × two sampling campaigns × three vineyards × three years), from which 54 started a spontaneous fermentation, 36 were not able to start fermentation after 30 days of incubation, whereas 18 samples were not collected due to unfavorable weather conditions and a bad sanitation state of the grapes in 2002. From the 54 fermentations 1620 yeast isolates were obtained. All the isolates were analyzed by their mtDNA RFLP (Hinf I) and a pattern profile was attributed to each isolate, resulting in a total of 297 different profiles.
The total yeast count (cfu in YPD medium) ranged between 1.0 × 106 and 8.0 × 107 cfu/ml must, corresponding to values generally described for grape must fermentations. All isolates belonged to the species S. cerevisiae due to their inability to grow in a medium containing lysine as sole nitrogen source and by their capacity to amplify six S. cerevisiae specific microsatellite loci. None or only 1 of the loci was amplified in other Saccharomyces species occurring in wine such as S. paradoxus and S. bayanus, respectively. No amplification was observed for species that are generally present at initial stages of fermentation, such as Candida stellata, Pichia membranifaciens and Kloeckera apiculata (not shown).
The results of mtDNA RFLP for the 1620 isolates are summarized in Table 1. Among the total 450 isolates collected in vineyard A, 93 corresponded to unique patterns whereas in C and P a total 450 and 690 strains were isolated, corresponding to 62 and 135 unique patterns, respectively.
Table 1. MtDNA RFLP analysis of 1620 yeast isolates from fermented must prepared with grapes collected in vineyards A, C and P of the Vinho Verde Region, indicated in Fig. 1, during the harvest of 2001, 2002 and 2003
| || ||Site||Number of isolates||Number of distinct patterns||Number of unique patterns||Common patterns|
|2001||E||AI AII AIII||NF||–||–||–|
| || ||AIV AV AVI|| || || || |
| ||L||AI AIV AVI||NF||–||–||–|
| || ||AII||90||2||10||A06|
| || ||AIII|| ||8|| || |
| || ||AV|| ||1|| || |
|2002||E||AI AII AIII||NC||–||–||–|
| || ||AIV AV AVI|| || || || |
| ||L||AI||180||16||34||ACP10 A06 A11|
| || ||AII|| ||2|| ||ACP10|
| || ||AIII|| ||9|| ||A11 A13|
| || ||AIV|| ||6|| ||A06 A13|
| || ||AV|| ||9|| ||A13|
| || ||AVI|| ||1|| ||–|
|2003||E||AI AII AIII||NF||–||–||–|
| || ||AIV AV AVI|| || || || |
| ||L||AI||180||3||41||ACP10 S3|
| || ||AII|| ||1|| ||ACP10 S3|
| || ||AIII|| ||9|| ||A06|
| || ||AIV|| ||12|| ||–|
| || ||AV|| ||19|| ||–|
| || ||AVI|| ||2|| ||–|
|2001||E||CI CII CIII||NF||–||–||–|
| || ||CIV||90||5||2||S1 S2 S3 S4 S5|
| || ||CV|| ||4|| ||S4 S5|
| || ||CVI|| ||1|| ||S1|
| || ||CII||150||20||24||S5|
| || ||CIII|| ||4|| ||S1 S4 S5|
| || ||CIV|| ||2|| ||S3 S5|
| || ||CV|| ||4|| ||S1 S3 S4 S5|
| || ||CVI|| ||8|| ||S1 S2 S4 S5|
|2002||E||CI CII CIII||NC||–||–||–|
| || ||CIV||30||1||1||–|
| || ||CV CVI||NF||–||–||–|
| ||L||CI CII CIII||NC||–||–||–|
| || ||CIV CV CVI|| || || || |
|2003||E||CI CIII CIV||NF||–||–||–|
| || ||CV CVI|| || || || |
| || ||CII||30||3||3||–|
| ||L||CI||180||8||32||S3 S4 C69P77 C63|
| || ||CII|| ||3|| ||C63|
| || ||CIII|| ||1|| ||ACP10|
| || ||CIV|| ||18|| ||S1, C42P80|
| || ||CV|| ||9|| ||–|
| || ||CVI|| ||2|| ||S4|
| || ||PII|| ||2|| ||P136|
| || ||PIII PIV PV PVI||NF||–||–||–|
| ||L||PI||180||6||62||S3 S5 S6 P136|
| || ||PII|| ||17|| ||S3 S6 ACP10 P03 P136|
| || ||PIII|| ||8|| ||–|
| || ||PIV|| ||21|| ||P03 P24 P50 P136|
| || ||PV|| ||15|| ||S3 P24|
| || ||PVI|| ||13|| ||S3 S6 P136|
|2002||E||PI PII PIII||NF||–||–||–|
| || ||PIV PV PVI||NC||–||–||–|
| || ||PI||150||5||12|| |
| || ||PII|| ||4|| ||ACP10 S6 P136|
| || ||PIII|| ||1|| ||ACP10|
| || ||PV|| ||10|| ||S3 S6 P50 P136|
| || ||PVI|| ||1|| ||ACP10|
| || ||PI||120||10||12||–|
| || ||PII|| ||1|| ||–|
| || ||PIII|| ||1|| ||P136|
| || ||PVI|| ||2|| ||ACP10|
| ||L||PI||180||15||47||ACP10, S3 C69P77|
| || ||PII|| ||1|| ||–|
| || ||PIII|| ||9|| ||P136|
| || ||PIV|| ||18|| ||S3|
| || ||PV|| ||5|| ||C42P80|
| || ||PVI|| ||5|| ||–|
For 11 common patterns, found in more than one fermentation (Table 1 and Fig. 2), and also for six commercial starter yeast strains (S1–S6), a wider geographical and temporal distribution was verified. Patterns S1–S6 corresponded to commercial starter yeasts that had been used in the wineries for the last few years. Perennial strains were associated with more sites of a single vineyard (patterns A06 and S6, P136, P50), but showed also a wider distribution across multiple sampling sites in two or three vineyards (patterns S3, S4, and ACP10). Patterns S1, S2, C63, A11, A13, P03 and P24 were found only in one year but across several sampling sites of a single vineyard, while strain S5 had a wider distribution across several sampling sites of vineyard C and P. Patterns C42P80 and C69P77 appeared only in a single sampling site during 2003 of both vineyards C and P. Pattern ACP10 is the only “regional” isolate with a wider geographical distribution, whereas A06, A11, A13, C63, P03, P27, P50 and P136 can be considered as “vineyard-strains” due to their occurrence in multiple sampling sites and/or years.
Figure 2. Examples of common mitochondrial DNA RFLP (Hinf I) patterns, as listed in Table 1, found in yeast strains isolated from spontaneous fermentations of must collected as described in Section 2.
Download figure to PowerPoint
The wet weather in the summer 2002 resulted in severe fungal infestations and heavy applications of chemical sprays. This was possibly the reason that merely 12 unique patterns were identified among the 150 strains collected in the late sampling stage in 2002 in vineyard P. In 2003, this relation was again more similar to the one found in 2001 (47 and 62 unique patterns among each 180 isolates from the late sampling stages of vineyard P).
As shown in Fig. 3, onset of spontaneous fermentation was verified in almost all grape samples collected in the late sampling campaign. This was rarely the case for most of the samples collected some days before the harvest. Must prepared from grapes collected in the early sampling stage in vineyard A, never started to ferment spontaneously. An accidental agrichemical over-dosage occurred in 2001, resulting in delayed spontaneous fermentation onset for three of the four post-harvest samples (II, III and VI). In the following two years, fermentation profiles were similar to samples from C and P, suggesting the recovery of the intervenient flora.
Fermentation started after six to twelve days being generally accomplished by one to twenty strains. Apparently no correlation between the number of strains involved in a fermentation and sampling site, year or vineyard was found. The wider distributed strain (ACP10) was dominant in six fermentations (AII-2002, AI-2003, AII-2003, CIII-2003, PIII-2002, PVI-2002) contributing to 77–100% (23–30 strains) of the total yeast flora, but was of minor importance in five fermentations (AI-2002, PII-2001, PII-2002, PI-2003, PVI-2003), accounting for only 3–10% (one to three strains), and being accompanied by one to sixteen different strains. The distribution of this strain is not associated with the capability to predominate in fermentation, and competition with accompanying strains seems to play the key role.
Vineyard-specific patterns of samples collected in the early stage did not appear after two weeks at the same site (P, 2001 and 2003, C, 2001) with the exception of the more generalized patterns S1, S2, S3, S4, S5, ACP10 and P136, speaking in favor of a very diversified S. cerevisiae flora.
- Top of page
- 2Materials and methods
Biogeographical large-scale surveys and studies on the genetic diversity of S. cerevisiae strains isolated from spontaneous fermentations have documented the dynamic nature of these populations. In the present study, 297 different genetic patterns have been found among 1620 isolates obtained from 54 small scale fermentations performed with grapes from three vineyards located in the Vinho Verde Region, during a three years period. The vast majority of the patterns were unique, demonstrating an enormous biodiversity of S. cerevisiae strains in the Vinho Verde Region. Considering the ratio between the number of isolates and the number of patterns as an approximate biodiversity estimative, our results showed similar values to previously published surveys on genetic diversity of autochthonous oenological S. cerevisiae strains in other regions with viticulture traditions such as Bordeaux, Charentes [17,45], Campagne and Loire Valley, in France; El Penedèz, Tarragona, Priorato [20,22] and La Rioja in Spain; Germany and Switzerland; Tuscany, Sicily and Collio in Italy; Amyndeon and Santorini in Greece; Western Cape [16,18,43] in South Africa; and Patagonia in Argentina.
The present study has been carried out in a viticultural region that has never been characterized before and includes aspects that have not been considered in previous works, such as the appearance of several commercial yeast strains, and the comparison of yeast populations that can be found in grape samples before and after the harvest.
The vast majority of the strains did not display a perennial behavior, being the flora of each year characterized by the appearance of many new patterns. This might be attributed to the sampling of only 12 × 2 kg of grapes per vineyard and year, being not enough to grasp the entire biodiversity wealth of a given area. Another reason for the appearance of new patterns could be attributed to recombination and evolutionary forces, but it seems unlikely that such changes occur from one year to another to justify the presence of many distinct patterns in consecutive years. Mitochondrial DNA RFLP patterns are stable when S. cerevisiae cells undergo about five to seven divisions during alcoholic fermentation (our unpublished data).
Among all patterns only ACP10 showed a wide regional distribution with a perennial behavior providing preliminary evidence for a strain representing a “terroir” as described [17,21]. However, the wider distribution of a strain is not necessarily correlated with a better technological fitness. This makes sense from an ecological point of view, since the selective forces that act in a vineyard are completely different from those that yeast may find in a fermenting grape must. Further physiological characterization under wine making conditions is required to evaluate the potentialities of this strain. The appearance of this strain did not obey to a generalized pattern, but rather to sporadic presence, absence and reappearance, due to natural population fluctuations. The perennial appearance of pattern ACP10 is a consequence of its prevalence in the local microflora. In different fermentations, ACP10 was dominant or not, showing that the final outcome of fermentation was dependent on the specific composition of the yeast community in the must, that is influenced by many factors such as the killer effect which depends strongly on the ratio of killer to sensitive cells at the beginning of the fermentation.
Grape variety of vine A was Loureiro, being Alvarinho and Avesso the cultivars of vineyard P and C, respectively, indicating that the grape variety could contribute to the finding of so many distinctive patterns. Traditional wine-making practices are very similar in A, C and P, and differences in climatic influences seem to be of minor influence since the three vineyards are geographically close. However, one can not exclude microclimatic influences, not recorded in the present study.
A first sampling campaign was performed some days before the harvest; a second was carried out a few days after the end of harvest. This was accomplished in a time frame of about two weeks, in order to obtain a more detailed picture of the temporal distribution of fermenting yeast populations during the harvest. As grapes mature to full ripeness, yeasts become more abundant. The last stage of the grape maturation can favor fermentative yeast proliferation on grape surfaces, due to the decrease of grape skin integrity and must leakage from the berries. Insects are the probable source of yeast on damaged grapes. Yeast colonization of grapes can reach values of about 105–106 cfu/berry. Before vintage, only 5% of the grapes harbor yeasts, while this number is much higher (60%) during vintage. As expected, only 11 of 42 pre-harvest samples (26%) were able to ferment spontaneously compared to 43 of 48 post-harvest samples (90%). The associated strains were also much more diversified in the late sampling campaign (267 patterns among 1260 isolates) compared to the early stage (30 patterns among 360 isolates). With only one exception (pattern P136), autochthonous strain patterns from the early sampling stage did not appear in the late sampling stage, speaking in favor of a succession of S. cerevisiae strains. Alternatively, differences can be attributed to the fact that different grape bunches were harvested, that may have, although in close proximity to each other, a distinct flora. It seems unlikely that the enormous increase in strain variability at harvest time is due to a spreading of winery-resident flora with harvesting equipment.
Spontaneous fermentations were performed by one or more predominating strains accompanied by no, few or many “secondary” strains, or by a very heterogeneous yeast community with no prevalent strain(s). This is in agreement with other studies reporting the presence of one or two predominating strains constituting more than 50% of total biomass, and a varying number of “secondary” strains [7,17,19,29,40,41], or presence of many distinct strains with no prevalence [22,42]. The occurrence of both situations has also been reported [16,18,43].
Considering that the origin of wine yeasts is still controversial [3,5,8,44], our results clearly indicate that S. cerevisiae occurs in vineyard ecosystems belonging to the Vinho Verde Region in sufficient high numbers to conduct a spontaneous fermentation from musts prepared with approximately two kg of grapes. However, some remarks have to be made concerning our experimental approach. Grape must creates selective and very stressful conditions for yeast, totally distinct from the environmental influences in nature. It is therefore clear that our data refer only to S. cerevisiae strains capable to survive the conditions imposed by fermentation, under our experimental circumstances, giving therefore a distorted picture (underestimation) of the kind of strains really occurring in vine. As the detection limit of our experimental approach is 3.3% (one strain in 30 isolates), rare strains, although capable to survive fermentation, might also have not been detected. Searching for S. cerevisiae in 18 sites, in two campaigns and over three years using a direct-plating method from single grape berries, as described would be highly labor-intensive. Therefore we regard our approach as an acceptable compromise, allowing good estimation of population composition, but preventing a precise description in terms of relative strain abundance in nature.
The present work is the first large-scale approach about the vineyard-associated strains from the Vinho Verde Region in Portugal, being a useful approach to obtain a deeper insight into ecology and biogeography of S. cerevisiae strains, even among geographically close regions. We consider these studies indispensable for the developing of strategies aiming at the preservation of biodiversity and genetic resources as a basis for further strain development.