OUP user menu

Peculiarities of flor strains adapted to Sardinian sherry-like wine ageing conditions

Marilena Budroni, Severino Zara, Giacomo Zara, Giorgia Pirino, Ilaria Mannazzu
DOI: http://dx.doi.org/10.1016/j.femsyr.2005.04.002 951-958 First published online: 1 July 2005


Saccharomyces cerevisiae flor yeasts, which are subjected to stressful conditions during wine ageing, exhibit a number of characteristics which distinguish them from non-flor S. cerevisiae wine strains. In the present work, 22 flor and 14 non-flor S. cerevisiae wine strains are compared, in order to elucidate other possible peculiarities of these yeasts. The results obtained demonstrate that in contrast to the homothallic nature of the non-flor strains, 77% of the flor strains exhibit two variants of a semi-homothallic life cycle. Moreover, the flor-forming ability is shown to be inversely correlated to spore viability and the utilisation of maltose and galactose.

  • Saccharomyces cerevisiae
  • Life cycle
  • Principal component analysis
  • flor
  • Acetaldehyde

1 Introduction

Different wine-making conditions can affect the genetic characteristics of wine yeast strains. These include the presence of polymorphisms at the chromosomal level [1] such as aneuploidy, polyploidy, duplications or deletions of chromosomal regions, and the presence of hybrid chromosomes [26]. Karyotype variability appears to be a consequence of the highly selective pressure exerted by the winery environment [6,7]. In particular, aneuploidy, which leads to genotypes that are deficient in sexual reproduction and which can be a cause of sexual isolation [8], is a consequence of the elevated ethanol concentration in wines. Moreover, wine strains are extremely variable in their sporulation ability and spore viability [9,10].

Sherry-like wine ageing conditions are particularly restrictive, due to the long-lasting exposure to high levels of ethanol and acetaldehyde, which are well-known mutagens and inhibitors of many metabolic activities [1114]. As a consequence, flor strains exhibit peculiarities that include high polymorphism of the mtDNA [14] and the presence of a 24-bp deletion in the ITS1 region [15,16]. This deletion distinguishes them from other wine yeasts. However, to date, little is known about the effects on the life cycle of Saccharomyces cerevisiae wine strains that are due to the winery environment.

Mortimer [17] has indicated that 69 % of S. cerevisiae wine strains are homothallic [18], and has proposed a “genome renewal” model for these strains [18]. This states that together with sporulation, homothallism represents a selective advantage for wine yeasts by providing a means for the evolution of advantageous character selection in the homozygous state [18]. About 10% of S. cerevisiae wine strains have a heterothallic life cycle [18] and in this case, haploid generations are not able to self-diploidise due to a mutation in the HO gene. Finally, some studies have highlighted the existence of a third type of life cycle in flor yeasts [1922]. This is a semi-homothallic cycle which is characterised by an anomalous segregation with two non-mating and two mating cultures (MATa or MATα).

To elucidate other possible peculiarities of these yeasts, 22 flor and 14 non-flor S. cerevisiae wine strains were subjected to genetic and phenotypic analyses and their life cycles have been studied. Here, we show that 77% of the flor strains have two variants of the semi-homothallic life cycle, and that flor-forming ability is inversely correlated to spore viability and the utilisation of maltose and galactose.

2 Materials and methods

2.1 Yeast strains

All the yeast strains used were isolated from wines and fermenting musts that were sampled in cellars, which had not been contaminated by commercial starter cultures. They thus represented the result of strong selective pressures that occur during different wine-making conditions. Strains V45, V97, V48, V42, V11, V44, V19, V75, M25, M26, M38, M39, M45, M46, A9, A28, A33, A43, A51, A56, A58 and A68 were isolated from sherry-like wines sampled in Sardinia and were able to form a flor. Strains 1184 and 1098 are non-flor-forming yeasts isolated from Sardinian wines. All these strains belong to the collection of Dipartimento di Scienze Ambientali Agrarie e Biotecnologie Agroalimentari (DiSAABA). Strains Sc12, Sc23, Sc28, Sc42, Sc44, Sc93, Sc112, Sc143, Sc151, Sc157, Sc200 and Sc205 were isolated from wines sampled in the Marche Region of Italy [23]. All of these yeasts were ascribed to S. cerevisiae according to Kurtzman and Fell [24]. Haploid YPH499 (MATa) and YPH500 (MATα) were used as tester strains [25].

2.2 Phenotypic and genetic analyses

The yeast strains were maintained on YPD (1% yeast extract, 2% peptone, 2% glucose, 2% agar). Briefly, the ability to ferment sucrose (SUC), raffinose (RAF), maltose (MAL), melibiose (MEL), galactose (GAL) and trehalose (TRE) was tested on YPD plates containing the corresponding sugar plus the pH indicator bromothymol blue (BTB), and which were adjusted to pH 7.5 after autoclaving. Fermentation-positive cultures turned the agar beneath the strain from blue to yellow, as described by Mortimer et al. [18]. Copper resistance (CUP) was tested on YPD plates containing 60 mg l−1 copper ions. Hydrogen sulphide production was assayed as the formation of reddish-brown colonies on BIGGY agar (Difco, Franklin Lakes, USA). Tests for biofilm (flor) formation were performed in 24-well microplates with Yeast Nitrogen Base (Difco) containing 4% ethanol as the sole carbon source and which were supplemented, where necessary, with nitrogen bases and amino acids at standard concentrations [26]. Galactose and maltose assimilation and fermentation tests were performed also on YNB (0.67%) supplemented with 2% GAL or MAL as sole carbon source and bromothymol blue as indicator of fermentation. Assimilations were carried out on solid medium (2% agar) and fermentations were performed in liquid medium in test tubes containing Durham tubes.

Sporulations were performed on solid (1.5% potassium acetate, 2% agar) or in liquid (1% potassium acetate) medium. Asci were dissected on SC plates (6.7 g l−1 Yeast Nitrogen Base, 2% glucose, 2% agar) with 1% sorbitol as osmotic stabiliser. Spores from the first and second meiotic generations were evaluated for their fermentative abilities, H2S production, flor formation and sporulation. The mating types of non-sporulating strains were tested by crossing them with the strains YPH499 and YPH500. Life cycle segregation was tested by crossing the mating F1 derivatives with YPH499 and YPH500 [22].

2.3 Statistical analyses

Principal-component analysis (PCA) of the data derived from the genetic analysis was performed using the Unscrambler software (Camo Inc., Corvallis, OR, USA). For this purpose, the raw data of the segregation ratios in the progeny of each strain were converted into a numerical format. A segregation ratio of 4:0 (indicating the presence of the dominant allele in all of the spore progeny) was represented by 4, a segregation ratio of 2:2 (indicating the presence of the dominant allele in half of the spore progeny) was represented by 2, and a segregation ratio of 0:4 (indicating the absence of the dominant allele in all of the spore progeny) was represented by 0. Segregation ratios of 1:3 and 3:1 were not observed.

3 Results

3.1 Genetic and phenotypic characterisation

Two different S. cerevisiae populations were analysed, namely 22 flor strains, and 14 non-flor strains, that were unable to form flor. The whole set of strains was subjected to genetic and phenotypic analysis, as described by Mortimer and co-workers [18]. The ability to sporulate, ferment sucrose, raffinose, maltose, melibiose and galactose, assimilate glycerol, produce flor and hydrogen sulphide, and tolerate copper were evaluated in single-spore clones derived from each strain.

Spore viability was generally lower in flor strains that than in non-flor strains as shown in Fig. 1. Spore viability of the flor strains did not increase after the addition of 1% sorbitol to the growth medium [27], and it was necessary to select a second meiotic generation of these strains to increase their spore viability (Fig. 2).

Figure 1

Spore viability of flor and non-flor yeasts. Sc indicates non-flor strains.

Figure 2

Spore viability of second meiotic generation of flor yeasts.

None of the flor and non-flor strains were able to ferment melibiose, similar to the observations made by Mortimer and colleagues [18]. Different segregation ratios were observed for all of the remaining characters, and the data obtained were subjected to PCA using the Unscrambler software. The 2D scattergram obtained for the 1st and 2nd principal components, which accounted for 64% of the variance (Fig. 3), provided interesting information regarding the existence of correlations between different characters. As shown in Fig. 3, the flor-forming ability was inversely correlated to maltose and galactose utilisation and the ability of individual spore clones to sporulate.

Figure 3

Principal-component analysis (PCA) of the data from the genetic analysis. The scattergram for the 1st and 2nd principal components, accounting for 38% and 26% of the variance, clustered the 37 isolates into two groups. The one on the left contains the whole set of flor strains, except V97, whereas the one on the right comprises the non-flor strains.

Indeed, single-spore clones produced from flor strains (except V97) were unable to ferment maltose and galactose. In contrast, the progeny of the non-flor strains (except Sc143) were able to ferment maltose and/or galactose. The ability to ferment raffinose and sucrose was variable among the strains analysed, as could be concluded from the distance to the axis. In contrast, copper resistance (CUP) and glycerol assimilation (GLY) were positioned very close to the axis origin, as could be expected from the low variability in the segregation ratio for these characters. Also shown in Fig. 3, and in accordance with the observed differences, the whole set of strains cluster into two groups: one comprising the flor strains (Fig. 3, left), and the other containing all of the non-flor strains and V97 (Fig. 3, right), which differed from the flor strains in spore viability and maltose and galactose fermentation. Galactose and maltose fermentation and assimilation abilities were further investigated in a larger wine yeast collection comprising 48 flor and 54 non-flor strains belonging to the DISAABA collection. Interestingly, 73% and 69% of the flor yeasts were unable to ferment and assimilate galactose and maltose, respectively, while the non-flor strains were generally able to ferment and/or assimilate galactose and maltose (Table 1).

View this table:
Table 1

Galactose and maltose fermentation and assimilation in flor and non-flor strains


3.2 Life cycle

All of the non-flor strains and four of the flor strains, namely (A28, V97, V42 and V11) were homothallic. The only heterothallic strain was the flor strain M39.

Of the 22 flor strains, 17 showed a semi-homothallic life cycle (Table 2). Ten of these, producing meiotic derivatives with a segregation of two mating: two non-mating (sporigenous), showed segregations Hq (two mating MATa and two non-mating) or Hp (two mating MATα and two non-mating) [20,21]. In particular, V75, M25, M45, V45, M26 and V44 showed a semi-homothallic cycle with Hq segregation. This was confirmed in the second meiotic generation for strains V75, M25 and M45 only. V48 showed MATa segregation only in F2, and the remaining three strains (V19, A9 and A33), showing a semi-homothallic cycle with Hp segregation, confirmed this sexual behaviour in F2 (Fig. 4).

View this table:
Table 2

The life cycles of the flor yeast strains studied

StrainsFirst meiotic generationSecond meiotic generationHybrid
Semi-homothallic cycle (Hq–Hp)
V75+/−a+/−a+/−a and α
M25+/−a+/−a+/−a and α
M45+/−a+/−a+/−a and α
V45+/−a+/++/−a and α
M26−/−a+/−a and α
V44+/−a+/++/−a and α
V48+/++/−a+/−a and α
V19+/−α +/−α +/−a and α
A9+/−α +/−α +/−a and α
A33+/−α +/−α +/−a and α
Semi-homothallic with bisexual meiotic derivatives
M46+/−a+α+/++/−a and α
A43+/−a+α+/++/−a and α
A58−/−a+α+/−a and α
A68+/−a+α+/++/−a and α
A56+/−a+α+/++/−a and α
A51+/−a+α+/++/−a and α
a and α
  • Spo: sporulation ability of the meiotic derivatives; MAT: mating type; Hq: segregation of two mating (MATα): two non-mating; Hp: segregation of two mating (MATa): two non-mating.

  • aSemi-homothallic cycle with bisexual meiotic derivatives even in the second meiotic generation of the hybrid.

Figure 4

The semi-homothallic cycle with Hp or Hq segregation. (a) Starting from different asci both sporigenous (SPO) and mating cultures (MATa in Hq and MATα in Hp) are obtained. Spore viability in F1 is generally low and produces both viable and dead spores. (b) Sporigenous cultures in F2 segregate as two sporigenous: two mating cultures (MATa in Hq and MATα in Hp). (c) Few of the cultures obtained from their hybrids can sporulate.

Strains M46, A43, A51, A56, A58, A68 and M38 showed another type of life cycle, that is indicated as semi-homothallic with bisexual derivatives, as in their first meiotic generation they produced bisexual and sporigenous derivatives. The presence of bisexual cultures increased the potential for crossing, as these are able to cross with MATa and MATα strains. The sporigenous derivatives showed a typical homothallic behaviour in F2, thus differing from sporigenous derivatives of strains harbouring a semi-homothallic life cycle with Hp or Hq segregation. None of these strains confirmed this behaviour in the second generation (Fig. 5).

Figure 5

The semi-homothallic cycle with bisexual meiotic derivatives. (a) Following meiosis both sporigenous (SPO) and bisexual cultures (MATa+ MATα) are obtained. (b) The sporigenous culture produce only sporigenous cultures in F2. (c) The bisexual culture is able to cross with strains of both (a or α) mating types. (d) Hybrids with heterothallic strains segregate as sporigenous cultures.

The crosses between F1 derivatives of semi-homothallic strains and YPH499 and YPH500 produced hybrids which gave spore clones with two different segregation ratios (Table 2). Interestingly, hybrids produced by mating the F1 derivatives of strains with a bisexual behaviour and YPH499 and YPH500 gave rise to spore clones showing both the segregation ratios observed in semi-homothallic strains with Hq and Hp segregations (Fig. 5, Table 2).

4 Discussion

S. cerevisiae flor strains differ from non-flor strains belonging to the same species in physiological (e.g., high alcohol resistance, acetaldehyde production, oxidative metabolism, film [flor] formation ability) and genetic features [10,15]. The results presented here revealed further peculiarities of the flor group. Spore viability, which is conditioned by genetic and environmental factors [28] and has been shown to vary from 0% to 98% in wine strains [11] was generally lower in flor strains, as compared to the non-flor strains. Low spore viability can be due to several causes such as chromosome loss after spore germination [27]. In the experimental conditions tested, low spore viability persisted after the addition of an osmotic stabiliser. Thus, this character appears to be due to abnormal chromosomal complements resulting from the loss, duplications or rearrangements of chromosomes, rather than to defects in the spore wall [27].

Another significant difference between the two populations analysed was the ability to ferment and/or assimilate galactose and maltose. This was confirmed in a total of 70 flor and 68 non-flor S. cerevisiae wine strains. Most of the non-flor strains were able to utilise galactose and maltose, while the great majority of the flor strains were unable to both ferment and assimilate these sugars. Interestingly, Johnston and collegues [28] have reported that the GALo phenotype (lack of galactose utilisation) is due to mutations of either the GAL7 or GAL10 genes, which are located on chromosome II and are close to at least five open reading frames (ORFs) that are prone to meiosis-induced double-strand breaks (DSBs) in Spanish flor strains [29]. The frequency of the GALo phenotype was significantly different between the two groups of strains, as it was present in 75% of the flor strains and only in 39% of the non-flor strains (Table 1).

Naumov and co-workers [30] have shown the existence of only two MAL loci (MAL1 and MAL3) in wine strains, which are located on chromosomes VII and II, respectively. The lack of maltose utilisation (the MALo phenotype) was observed in 9% of the non-flor and in 73% of the flor strains (Table 1).

Considering that flor yeasts produce high concentrations of acetaldehyde during wine ageing, and that this is one of the major causes of DSBs and other major chromosomal rearrangements [11], it can be hypothesised that the GALo and MALo phenotypes are caused by DNA damage due to high acetaldehyde concentrations. Similarly, the low spore viability observed in the flor strains may be due to abnormal chromosomal complements.

While all of the non-flor strains were homothallic, the flor strains showed a great heterogeneity in their life cycles. Indeed, only four of the flor strains were homothallic, one was heterothallic, and the remaining 17 were semi-homothallic with Hp/Hq segregation or bisexual derivatives.

According to Mortimer et al. [18] a homothallic life cycle, which occurs most frequently in wine strains [17], provides adaptive advantages due to the occurrence of diploidisation after mating type inter-conversion. This allows the elimination of unfit characters and the evolution of the strain through genome renewal [18,31]. Heterothallism occurs much less frequently in wine strains (10%), probably because heterothallic strains are prone to the negative effects of dominant unfit mutations at the heterozygous state and, therefore, they are less adaptable to the stressful conditions that occur during wine making [17].

The question arises whether the semi-homothallic life cycle is a consequence of or provides a better adaptation to the restrictive environmental conditions typical of sherry-like wine ageing.

Fifty percent of the meiotic derivatives of semi-homothallic strains are sporigenous and show homothallic behaviour. They represent a conservative “choice” that allows genome renewal, as unfit heterozygous characters can be eliminated, while the characters that allow the strain to survive under such selective wine making conditions are maintained. In the case of semi-homothallic strains with Hp/Hq segregation, the other 50% meiotic derivatives are mating clones of one sexual type (either a or α). Thus, Hp and Hq segregation, by allowing mating, tend to increase heterozygosis, which can have positive and/or negative consequences for the survival of these strains under selective conditions [22], such as during wine fermentation and ageing. The strains showing a semi-homothallic life cycle with bisexual derivatives follow the same pattern as seen for the evolution of the semi-homothallic strains with Hp/Hq segregation, although the presence of bisexual cultures increases the possibility of crosses with both the MATa and MATα strains. Crosses between strains of opposite mating types are necessary to acquire diploidy, and they allow the hybrid progeny to recombine the parental characters and to increase their potential for survival. As shown by Budroni et al. [22], the mating progeny of the semi-homothallic strains has a dominant HO gene.

We observed that the flor strains have low spore viability can contribute to strain genetic stability [17]. Interestingly, the sporulation of strains M38, M26 and A58 produced exclusively mating derivatives in F1 [32], while the increase in spore viability in F2 allowed these strains to produce sporigenous and mating derivatives. Considering that sexual crossing increases the average fitness in environments to which the populations are well adapted [33], this suggests that a semi-homothallic life cycle provides two chances of adaptation: one through genome renewal, and the other through hybridisation.

Another interesting feature of the semi-homothallic flor strains is that they can be used for genetic improvement (M. Budroni, unpublished data). By means of cell-to-cell crosses, they can be used to produce hybrids able to generate homothallic cultures that can offer the following advantages: (i) the availability of mating cultures, which can be characterised before crossing and compared with cultures that undergo autodiploidisation; (ii) the crossing can be performed in plates without the need for micromanipulation; and (iii) the production of derivatives for genetic studies. Moreover, these strains can be used as further biological models for the study of the dynamics and the genetic bases of adaptation.


The authors are grateful to Giovanni Antonio Farris for discussions and comments on the manuscript. We thank Giovanni Giordano, Fabio Tavera and Monica Sechi for their assistance and helpful suggestions, Emanuele Boselli for help with the statistical analysis of the data and Gianfranca Ladu for performing galactose and maltose fermentation tests. This research was supported by PON-Misura 1.3: progetto AgriEchnos.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
View Abstract