Flor yeasts are a particular kind of Saccharomyces cerevisiae strains involved in Sherry wine biological ageing. During this process, yeasts form a film on the wine surface and use ethanol as a carbon source, producing acetaldehyde as a by-product. Acetaldehyde induces BTN2 transcription in laboratory strains. Btn2p is involved in the control of the subcellular localization of different proteins. The BTN2 gene shows a complex expression pattern in wine yeast, increasing its expression by acetaldehyde, but repressing it by ethanol. A flor yeast strain transcribes more BTN2 than a first fermentation yeast during growth, but less under different stress conditions. BTN2 deletion decreases flor yeast resistance to high ethanol concentrations. Surprisingly, this effect is suppressed by the addition of high amounts of amino acids to the growth medium, indicating that the role of Btn2p protein in amino acid transport is important for ethanol resistance. Btn2p deletion increases the fermentative capacity of flor yeast and its overexpression prevents its growth on nonfermentable carbon sources. BTN2 deletion also affects the biofilm formation ability of flor yeast, and it increases its sliding motility, resulting in increased mat formation. This correlates with an increased transcription of the FLO11 gene, a gene essential for biofilm formation.
The Btn2p protein is a v-snare interacting protein involved in intracellular protein trafficking (Kama et al., 2007). It facilitates specific protein retrieval from the late endosome to the Golgi apparatus. Consistent with this role in protein localization, its deletion caused mislocalization of several proteins. First, it was described that the mutant failed to localize the Rhb1p protein to its particular cell periphery localization (Chattopadhyay & Pearce et al., 2002). Rhb1p regulates the activity of the arginine and lysine transporter Can1p (Urano et al., 2000). Next it was shown that Yif1p changes its localization from the Golgi apparatus to the vacuole in the btn2Δ deletion mutant (Chattopadhyay et al., 2003). Finally, it was reported that Btn2p is necessary for the right localization of the salt tolerance-involved plasma membrane Ist2p (Kim et al., 2005). Btn2p physically interacts with all these proteins; thus, it seems to be a key regulator of their traffic.
The multiplicity of Btn2p interactions may explain the pleiotrophic phenotypes of the btn2Δ deletion strain, which has been involved in ion homeostasis, amino acid transport and pH regulation (Kim et al., 2005). Its regulatory role in the arginine plasma membrane carrier Can1p via Rhb1p results in a defect on arginine uptake in the mutant strain (Chattopadhyay & Pearce et al., 2002). Besides, Btn2p is related to the vacuole arginine transporter Btn1p (Pearce et al., 1999). btn1 deletion results in the overexpression of BTN2, and btn2 deletion causes and increases vacuolar H+-ATPase activity (Chattopadhyay et al., 2000).
Flor yeasts are Saccharomyces cerevisiae strains involved in the manufacture of the Sherry fino type of wines (Jackson et al., 1994; Martínez et al., 1995; Infante et al., 2003). Winemaking involves the alcoholic fermentation of the sugars present in grape juice, mainly by S. cerevisiae strains, which we will call first fermentation yeasts from this point onwards. Ethanol is added to the dry wine up to a final concentration of 15–15.5% (v/v), and a film of yeasts develops on the surface. The so-called flor yeasts are metabolically and physiologically different from the first fermentation yeasts. Their metabolism is basically respiratory, and they use the ethanol and glycerol produced during fermentation as carbon sources. In this process, acetaldehyde accumulates as a by-product of the oxidation of ethanol to acetate (Aranda & del Olmo et al., 2003). Flor yeasts show a strong resistance to the stress conditions present in the biological ageing of Sherry wines, i.e. high ethanol and acetaldehyde concentrations (Aranda et al., 2002).
The most distinctive characteristic of flor yeasts is their ability to float and form a buoyant film on the wine surface called velum. The key factor of this kind of growth lies in the proteins of the yeast surface (Martínez et al., 1997). Among these proteins, Flo11p/Muc1, a glycosylphosphatidylinositol-anchored cell wall protein, has been involved in velum formation on flor yeasts (Zara et al., 2005). In laboratory strains, Flo11p has been linked to both adhesion to plastic and sliding motility that cause growth on mats, i.e. complex multicellular structures composed of yeast-form cells (Reynolds & Fink et al., 2001). The ability of flor yeasts to form a biofilm has been related to an increased FLO11 expression because of a deletion in the FLO11 promoter in these strains (Fidalgo et al., 2006).
The BTN2 expression measured by microarrays is not increased by a 7% ethanol shock in laboratory strains (Alexandre et al., 2001). In fact, two of the three replicates of this global analysis show a decreased BTN2 transcription. Besides, genome-wide screenings of deletion mutants do not present a growth defect of the btn2 deletion mutant on ethanol (Fujita et al., 2006; van Voorst et al., 2006). However, the BTN2 expression is induced in laboratory strains by acetaldehyde, an abundant metabolite during fino biological ageing (Aranda & del Olmo et al., 2004), and it is highly transcribed by flor strains (Infante et al., 2002), suggesting that it could be relevant for ethanol stress in industrial strains and conditions. We carried out the deletion and overexpression of this gene on industrial wine yeasts in order to study its role in stress resistance and biofilm formation. Deletion of the BTN2 gene in flor yeast leads to reduced ethanol resistance and increased sliding motility. This indicates that Btn2p may play different and important roles for yeast performance during winemaking.
Materials and methods
Yeast strains and growth conditions
Saccharomyces cerevisiae flor strain 11.3 (Mesa et al., 2000) and first fermentation strain Fermivin (Gist-brocades) were used.
Yeasts were grown at 30 °C with shaking on liquid medium, and growth was followed measuring the OD600 nm. Rich media contained yeast extract 1.0% (p/v), Bacto peptone 2.0% (p/v) and one of the following carbon sources: glucose 2.0% (p/v) in the yeast potato dextrose (YPD) medium, ethanol 2.0% (v/v) in the yeast potato ethanol (YPEtOH) medium and glucose 10% (p/v) plus fructose 10% (p/v) in the yeast potato fructose glucose (YPFG) medium. In the YPD medium at a low pH, 25 mM tartaric acid was added and the pH was adjusted to 3.4.
Minimal medium SD contained yeast nitrogen base (YNB) 0.17% (p/v), (NH4)2SO4 0.5% (p/v) and glucose 2.0% (p/v). Amino acids and urea were added to a final concentration of 2.3 mM in the ethanol stress experiments. The flor medium was YNB 0.17% (p/v), (NH4)2SO4 0.5% (p/v) and ethanol 4.0% (v/v) (Zara et al., 2005). White wine contained ethanol 11% (v/v) and was provided by Bodegas Monteviejo.
The solid medium contained 2% agar, except the plates used for mat formation, which contained 0.3% agar. The YPD plates used for geneticine selection contained geneticine (G418, Gibco BRL) up to 200 mg L−1, whereas those used for canavanine resistance contained 60 μg mL−1 of this antibiotic. The SD plates with cycloheximide contained 2 μg mL−1 of the antibiotic.
The stress conditions were tested on cultures growing exponentially at an OD600 nm of 0.5. Ethanol was added to the culture up to a final concentration of 15% (v/v), and cells were incubated at 30 °C for 15 min to test the gene expression, and for 30 min to test viability. For the gene expression analysis at a low pH under YPFG conditions, cells were collected and incubated in these media at 30 °C for 15 min. Stationary phase cells were collected after 1 day of growth on YPD. Incubation in wine was carried out for 30 min at 20 °C. In the acetaldehyde stress measurements, cultures were preincubated for 1 h with 12% ethanol, acetaldehyde was added at 1 g L−1 and incubation was continued for 30 min. After the stress, viability was measured using a methylene blue solution (Sigma Chemical Co., St Louis, MO) on sodium citrate (2% w/v) at a final concentration of 0.01% (w/v) (Smart et al., 1999).
Yeast genetic manipulations
BTN2 deletions were carried out using the method developed by Güldener (1996). Briefly, the kanMX marker flanked by sequences homologous to BTN2 was amplified using oligos S1 (CGCGTTATTTTGATTGCAGTTCTCCAGTGAGCTATT ATCCCGTACGCTGCAGGTCGAC) and C2 (TCCTCAATAATAGAGTTTCCCTT GGGAGATCTGCTTAGGATAGGCCACTAGTGGATCTG).
Yeasts were transformed with this cassette using the lithium acetate method (Gietz & Woods et al., 2002), and geneticine-resistant colonies were isolated. Right transformants were selected by PCR and were transformed again with plasmid YEp351-cre-cyh to eliminate the kanMX according to Delneri (2000). Transformants were selected for cycloheximide resistance on SD plates, and the marker was eliminated inducing the Cre recombinase on YP+2% galactose. Oligos C2 (ATGTAAAAGGAGGAAAGCAATAAAAGCT AACCTAAGAGCTCGTACGCTGCAGGTCGAC) and S3 (GACTCGTTGTATCTGTCAACTTCCTCATCCTCAACGTCTTATAGGCCACTAGTGGATCTG) were used for the second and third deletions of the BTN2 gene in the 11.3 strain. These oligos hybridize to a BTN2 sequence that was already deleted on the first deletion mutant.
BTN2 overexpression was carried out using the strategy described by Cardona (2007) (see Fig. 2b). The SPI1 promoter-KanMX marker cassette was amplified using oligos BTN2a (GGAAAATTGCGACACTCGGTGGAGCTCGAGAGTTGTATCCTTCGTA CGCTGCAGGTCGAC) and BTN2b (GCAGCTGTTCAAAAACACATGGTGAA TTGAATATGGAAAACATTATTAGTAATAGTACTG) from the pKanMXSPI1p plasmid, and geneticine-resistant colonies were isolated.
Yeast fermentative capacity
The fermentative capacity was measured on YPFG medium. Two hundred and fifty millilitre bottles were filled with 100 mL of medium, and 107 cells mL−1 were inoculated from a preculture grown overnight in YPD. Cells were counted using a microtitre chamber. Cultures were shaken at 65 r.p.m. in a water bath at 30 °C. CO2 production was measured every 20 min using a Chittick device (American Association of Cereal Chemists). The fermentative capacity was expressed as millilitres of CO2 produced per number of cells over time.
Gene expression analysis
RNA was isolated using the hot-phenol procedure described by Kohrer & Domdey (1991). It was quantified by OD260 nm, and its quality and concentration were checked in 1% agarose gel in TAE buffer.
RNA analysis by Northern blot, hybridization and quantification were carried out as described previously (Aranda et al., 2002). The FLO11 probe was obtained by PCR using oligos AAGTTCTGCAAGCGCAGGCG and GGAATAGCCGTCGTGACAGG.
Real-time PCR was performed in a LightCycler 2.0. (Roche), using LightCycler® FastStart DNA MasterPLUS SYBR Green I according to the manufacturer's instructions. After 10 min at 95 °C, 40 cycles of 95 °C for 10 s, 54 °C for 3 s and 72 °C for 5 s were carried out using oligos GAGGATGAGGAAGTTGACAG and CCTTGGGAGATCTGCTTAG from BTN2. The amount of mRNA was calculated using the formula e[n−b/ax], considering n the crossing point, the number of cycles where the sample's fluorescence turns sharply upwards, and the rest of terms are derived from the calibration curve (y=ax+b), obtained measuring serial dilutions of 1 μL of the cDNA reaction for each gene studied. All the expression levels were normalized to that of the actine gene ACT1 using oligos CATGTTCCCAGGTATTGCCG and GCCAAAGCGGTGATTTCCT. The mRNA relative levels are the ratio to the ACT1 control, considering the levels of 11.3 strain growing exponentially in YPD as 1. All reactions were performed in triplicate.
Yeast adherence to plastic
Adherence was measured according to Purevdorj-Gage (2007). Cells were grown exponentially in YPD until they reached an OD600 nm of 0.8. Cells were washed and resuspended in YPD. Two hundred microlitres were transferred to a polystyrene 96-well microtitre plate, incubated at different times, washed with water and cells were stained with a 1% crystal violet solution for 15 min. After washing three times with 95% ethanol, the absorbance at A595 nm was measured with a Power Wave Ht by BioTek. A595 nm at time 0 was subtracted from A595 nm at different times, and the increment was represented over time.
BTN2 shows a complex transcriptional regulation in wine yeasts
To gain a better understanding of the expression of the BTN2 gene in wine yeasts, we analysed the mRNA levels under different conditions in both the 11.3 flor strain and the commercial wine yeast Fermivin (Fig. 1). BTN2 was highly expressed in both strains growing exponentially in rich medium YPD, although mRNA levels were higher in the flor yeast (Fig. 1a). When the cultures reached the stationary phase, both strains had similar BTN2 mRNA levels, and both were significantly lower than those detected during exponential growth. This behaviour is in contrast to that observed by a global analysis on a laboratory strain (Gasch et al., 2000). In these analyses, BTN2 was induced in the stationary phase, which reinforces the idea that transcriptional regulation may be different between laboratory and industrial strains under certain growth conditions. When both strains were transferred to the YPFG medium, a medium similar to YPD but with a higher sugar concentration designed to measure the fermentative capacity, BTN2 was repressed in both strains, but to a higher extent in the first fermentation strain Fermivin. The high amount of sugars in this medium may cause osmotic stress. In laboratory strains, osmotic stress causes a repression of BTN2 (Gasch et al., 2000).
The more relevant stress factors during Sherry wine biological ageing are ethanol and acetaldehyde (Aranda et al., 2002). Ethanol causes a decline in the BTN2 mRNA levels in both strains (Fig. 1a), but this decrease is greater in the flor strain. However, the effect caused by acetaldehyde in BTN2 is exactly the opposite (Fig. 1b). This is the only stress condition that causes an induction of this gene in industrial strains, and it is the same result that we found previously in a laboratory strain, where acetaldehyde triggers a major burst in BTN2 transcription (Aranda & del Olmo et al., 2004). Interestingly, this induction is higher in Fermivin than in the 11.3 strain. This may indicate that the flor strain is less affected at the transcriptional level by acetaldehyde.
Finally, we analysed the levels of BTN2 when both strains were transferred to white wine (Fig. 1a). This causes a decline of BTN2 levels in the Fermivin strain, but the flor strain maintains the original levels, indicating that it is not affected by the new environment. Levels in the Fermivin strain are similar to those seen in ethanol stress, and so the presence of ethanol in the wine may be the cause of the BTN2 repression. However, the high repression caused by ethanol in the flor strain was not observed when it was grown in wine, indicating that wine offers the flor strain an additional stimulus that compensates its ethanol content in terms of the BTN2 expression. This stimulus is not the pH; the decline in the pH to levels similar to those found in wine and grape juice (around 3.4) causes a repression of BTN2 in both kinds of wine yeasts (Fig. 1a).
BTN2 manipulation affects growth differently in different carbon sources
For a better understanding of the role of the BTN2 gene, we carried out two kinds of genetic manipulation on flor 11.3 and Fermivin strains (Fig. 2). Firstly, we deleted the BTN2 gene using the recyclable kanMX marker (Güldener et al., 1996). We repeated the deletion process three times in 11.3 and twice in Fermivin; hence, we assume these strains are at least diploid for this particular gene. Secondly, we constructed overexpressing strains by placing one of the copies of the BTN2 gene under the control of the stress-responsive promoter SPI1. To do this, we used a promoter exchange methodology, which was recently developed in our laboratory (Cardona et al., 2007). SPI1 is a nutrient-sensitive gene that is highly expressed in the stationary phase during wine fermentation (Puig & Pérez-Ortín et al., 2000), and under many different stress conditions (Gasch et al., 2000). We named the overexpressing strains 11.3 SB and FSB, respectively. The BTN2 expression increases up to ten times under stationary conditions in the modified strains (data not shown).
We tested the growth of these mutants in different media. In the rich medium YPD, the 11.3 btn2Δ deletion strain grows slightly worse than the parental strain at the beginning of the growth curve, although both reach a similar OD in the stationary phase. The overexpressing strain 11.3 SB has a consistently poorer growth and reaches a lower OD in the stationary phase (Fig. 3a). In another rich medium with a higher amount of sugar (YPFG), which we used to measure the fermentative capacity and that which mimics the grape juice sugar content, the pattern is the same but the differences are greater (Fig. 3b). The deletion causes a slight delay in growth that is more acute in the overexpressing strain. In a nonfermentable carbon source such as ethanol, the pattern differs (Fig. 3c). The overexpression of BTN2 is highly deleterious for growth, indicating that it plays an important role in respiratory growth in this flor yeast. Moreover, the deletion strain has a shorter lag phase and starts growing before the parental strain.
We also carried out the same genetic modifications in the first fermentation strain Fermivin, and growth in rich medium YPD was followed (Fig. 3d). Overexpression did not cause a significant effect on growth, but deletion led to slightly better growth during the first hours, although the final OD600 nm was the same. Growth on ethanol did not show any significant difference (data not shown). Therefore, Btn2p plays a different role in different strains and under different growth conditions.
BTN2 is necessary for tolerance to ethanol
Next, we tested the relevance of the BTN2 gene to resistance to high ethanol concentrations, the main stress condition that flor yeasts face during biological ageing. Flor yeasts tend to aggregate, and this phenotype makes it difficult to measure variations in the CFU number; hence, we used methylene blue staining (Smart et al., 1999) to determine cell viability. In the rich medium YPD, the deletion of BTN2 in flor yeast increases sensitivity to the high ethanol concentration (15% v/v; Fig. 4a) in both flor and first fermentation strains, particularly in the former (Fig. 4a). This is the reason why we carried out the next experiments with the flor strain. On an average, flor yeasts tend to be more resistant to ethanol than first fermentation strains (Aranda et al., 2002), but Fermivin shows a particularly high tolerance to ethanol; thus, it is useful to compare its behaviour to that of the flor yeast. The overexpressing flor strain 11.3 SB has a phenotype similar to the parental strain. However, the overexpressing Fermivin strain FSB has a slightly higher resistance to ethanol. A typical ethanol concentration during the biological ageing of Sherry wines is 15% ethanol. When a lower amount of ethanol (e.g. 11%) was used, the BTN2 deletion in flor yeast had no effect (Fig. 4b). Thus, Btn2p is a relevant protein to deal with high ethanol concentrations.
In biological ageing, acetaldehyde accumulates and results in a relevant stress condition. Flor yeasts are highly resistant to acetaldehyde alone, and so its effect on yeast viability is only seen in the presence of ethanol and resembles the biological ageing process (Aranda et al., 2002). We tested the resistance of the flor strain derivatives to 1 g L−1 acetaldehyde in the presence of 12% ethanol (Fig. 4c). Under this condition, the btn2Δ mutant shows a minimally increased sensitivity to acetaldehyde, and the overexpression of BTN2 displays a better survival rate, indicating a positive role of Btn2p in the tolerance to both simultaneous stress conditions. The BTN2 deletion mutant is not more sensitive to hyperosmotic and low pH stresses (data not shown); therefore, it does not show nonspecific stress sensitivity.
Btn2p is involved in arginine transport (Chattopadhyay & Pearce et al., 2002), and its deletion causes increased sensitivity to the arginine toxic analogue canavanine. The same is true when BTN2 is deleted in the flor strain (Fig. 4d). We wanted to test whether this phenotype is related to ethanol resistance. Therefore, we carried out the same experiments in the flor strain in the minimal medium to test the influence of amino acid composition on ethanol resistance. In the SD medium, the difference in ethanol resistance between the wild-type strain and the deletion mutant was even clearer (Fig. 4e). To test the role of arginine in this process, we added an excess of arginine to the medium. The arginine excess suppresses the difference in viability between wild-type and mutant strains under ethanol stress (Fig. 4d), indicating that the effect of Btn2p may be due to its role in regulating amino acid traffic. The addition of arginine decreases the ethanol tolerance in the wild-type strain, and it may be argued that arginine lowers ethanol tolerance. However, an excess of other amino acids such as lysine and glutamate also bypasses the need for Btn2p for ethanol tolerance without altering the wild-type behaviour to a great extent. These results indicate that Btn2p may not only control arginine traffic but also general amino acid transport. The addition of urea, a product of arginine catabolism, has no effect on the tolerance of both wild-type and btn2Δ strains (Fig. 4e), which also indicates that the presence of the whole amino acid is that which causes the effect, and not just the presence of a high concentration of a noncharged nitrogen source.
Finally, we tested the role of the deletion of BTN2 in the natural environment of the flor yeast, the white wine, and we found a lower resistance to ethanol in the mutant strain (Fig. 4f), which confirms the role of Btn2p in ethanol resistance under many different conditions.
Btn2p alters the fermentative capacity of flor yeast
To further study the effect of the BTN2 gene manipulation on yeast metabolism, we tested the fermentative capacity of the genetically modified flor yeast strains. The fermentative capacity was measured as CO2 production on a rich medium YPFG, a medium similar to YPD that contains a high amount of sugars, both glucose and fructose, which resembles grape juice. Surprisingly, BTN2 deletion increases the fermentative capacity of flor strain 11.3 (Fig. 5a). Its overexpression causes a slower carbon dioxide production.
Strain 11.3 is not devoted to fermentation and its metabolism inclines towards respiration under biofilm conditions. Therefore, we tested the same manipulations on a fully fermentative wine strain, Fermivin (Fig. 5b). In this case, both manipulations decreased the fermentative capacity of this strain, where the deletion was slightly more deleterious in terms of carbon dioxide production. Therefore, the metabolisms of both the flor and the first fermentation strains are quite different, but both are affected by the deletion of the BTN2 gene, which indicates that this gene is relevant in controlling carbon metabolism.
Btn2p plays a role in both mat formation and adherence to plastic
The main difference between flor yeasts and other wine yeasts is their ability to float on the wine surface and to form a film called velum. We tested whether the genetic manipulations carried out on the 11.3 flor strain changed the ability to perform this kind of growth. Yeasts grown overnight in rich YPD medium were inoculated into a minimal medium containing 4% ethanol, and incubated under static conditions (Zara et al., 2005). Both the wild-type strain and the btn2Δ deletion mutant formed a clear velum after 1 week (Fig. 6a). The overexpressing mutant was unable to form velum at all. Similar results were obtained when white wine was used as a growth medium (data not shown).
Velum formation is a complex event with many factors such as floatability, cell-to-cell interactions and ethanol tolerance. To investigate in more detail the effect of Btn2p in the relationship of yeast cells with their environment, we tested two parameters that indicate the status of the cell wall, such as adherence to plastic and mat formation (Reynolds & Fink et al., 2001). We carried out assays in microtitre plates to determine adherence to plastic. Samples of a culture were placed in the wells and were washed at different times (Fig. 6b). BTN2 deletion causes the ability to adhere to plastic to not increase with time, which occurs in the wild-type strain. After 4 h, the adherence to plastic of the mutant was lower than in the parental strain. Its overexpression causes the opposite effect, and the adherence to plastic is consistently greater in this mutant than in the other two strains. Therefore, Btn2p is needed to achieve a good adherence to surfaces.
Another way to test the interaction of the yeast surface with the medium is to follow mat formation in plates containing 0.3% agar, a measurement of sliding motility. Under this growth condition, the overexpression of BTN2 is detrimental for mat growth (see Fig. 6c). However, the deletion strain forms a consistently larger mat (because of the irregular form of growth, we could not measure the diameter of the mat, but the deletion showed larger mats in two sets of triplicates carried out independently). Therefore, Btn2p seems to control mat formation.
Both phenotypes are linked to the Flo11p/Muc1p protein, a cell surface glycoprotein (Reynolds & Fink et al., 2001). We checked the expression of the FLO11 gene in the flor yeast 11.3 in rich medium YPD in exponential growth and under stationary conditions (Fig. 6d). There is no detectable FLO11 mRNA in exponentially growing cells. The experiments of adherence to plastic were carried out under these conditions; hence, these results may indicate that Flo11p is not essential for adhesion in this yeast strain in exponential growth. Under stationary phase conditions, however, FLO11 is highly expressed, and is more expressed in the btn2 mutant (Fig. 6d and e). Therefore, it seems that the deletion of BTN2 causes an increased FLO11 expression in cells in the stationary phase that improves growth under mat-forming conditions. Its overexpression does not cause a significant change in the FLO11 expression (Fig. 6e), but it may cause additional growth problems that result in a poorer ability for mat formation (Fig. 6c).
Btn2p is a v-snare interacting protein involved in protein traffic in the late endosome to Golgi retrieval (Kama et al., 2007). The variety of cargoes and destinations (see Introduction) indicates that Btn2p is a key regulator of intracellular protein traffic, and explains the pleiotrophic effects of its mutation, which range from pH homeostasis, amino acid transport and tolerance to salt (Kim et al., 2005).
To those phenotypes, we add its role in ethanol resistance and biofilm formation. BTN2 deletion decreased ethanol tolerance in both flor (11.3) and first fermentation (Fermivin) industrial strains. It also causes higher sensitivity to ethanol in a laboratory strain (data not shown). If we consider its involvement in salt resistance, Btn2p may have a global role in stress tolerance. In laboratory strains and under certain environmental insults such as heat shock and oxidative stress, the BTN2 gene expression increases in a manner similar to the typical stress-responsive genes (Gasch et al., 2000). We detected a high expression of the BTN2 gene in the presence of acetaldehyde in a laboratory strain (Aranda & del Olmo et al., 2004). In both flor and first fermentation wine yeasts, the BTN2 transcription is also increased by acetaldehyde (Fig. 1b). However, the BTN2 expression decreases after ethanol, high osmolarity and low pH stresses in both kinds of wine yeasts. It is distressing that Btn2p is relevant to ethanol tolerance, but its gene is repressed under ethanol stress. This may indicate that the Btn2p effect on ethanol tolerance may be indirect or may occur at the posttranscriptional level. However, this seems to be the case in many genes related to ethanol resistance (discussed in Fujita et al., 2006). For instance, the deletion of genes of the vacuolar H+-ATPase led to ethanol sensitivity, but their expression levels were mostly downregulated or invariable in response to alcohols. Besides, BTN2 expression is activated by the presence of acetaldehyde, a metabolite derived from ethanol. This may indicate a tight transcriptional control of the BTN2 expression by both two-carbon metabolites. Under respiratory conditions, such as those that are present in the biological ageing of Sherry wines, a large amount of ethanol is metabolized to produce a certain amount of acetaldehyde. Under such conditions, the concentrations of both antagonistic metabolites in terms of the BTN2 expression may regulate the exact amount of Btn2p needed on each occasion. If this fine-tuning of the BTN2 expression is altered, the results are deleterious to the cell, particularly under respiratory conditions. In general, the overexpression of BTN2 under the control of a strong stress promoter, like SPI1, in flor yeasts causes a growth defect in most conditions, such as mat formation (see Fig. 6c). Nonetheless, it becomes dramatic when strains are grown under respiratory conditions like a rich medium containing 2% ethanol (Fig. 3c), and a minimal medium containing 4% ethanol (Fig. 6a).
It is also interesting to observe that the BTN2 deletion and overexpression, respectively, increase and decrease the fermentative capacity of the flor strain (Fig. 5a). In a first fermentation strain, metabolically different from the flor strain, both manipulations decrease the fermentative capacity (Fig. 5b). These circumstances indicate an important and complex role of Btn2p in the energy metabolism, which may affect the change from fermentative to respiratory growth. It is interesting that the BTN2 expression is reduced in wine strains at a low pH (Fig. 1a), because this protein is involved in pH homeostasis (Chattopadhyay et al., 2000). Because deletion in the Batten disease-related protein Btn1p increases the BTN2 expression, perhaps the BTN2 repression may similarly affect Btn1p and its role in pH homeostasis.
We believe that the role of Btn2p in amino acid transport may be linked to ethanol resistance. Accordingly, an excess of arginine, an amino acid whose transport depends on the Btn2p action, suppresses the phenotype of the btn2 deletion. We found that another positively charged amino acid (lysine) had the same effect, and surprisingly, an unrelated amino acid (glutamate) did the same. These results indicate that amino acid homeostasis is a key factor in regulating ethanol resistance. This may be related to the fact that an accumulation of proline in the vacuole also leads to ethanol resistance (Takagi et al., 2005). The homeostasis of ionic species, such as amino acids, may be a mechanism used by S. cerevisiae to control membrane potentials and ionic pump activities. The addition of an uncharged nitrogen source such as urea does not suppress the btn2Δ phenotype, thus reinforcing this hypothesis. We think that increased sensitivity to ethanol of the btn2Δ strain may be related to the failure to localize one of its cargo proteins, or several, to their right location.
A very relevant phenotype from the point of view of flor yeasts and their ability to form biofilm is the increased transcription of the FLO11 gene in the btn2Δ mutant, which results in an increased rate of mat formation. We do not believe that Btn2p is directly involved in transcriptional regulation. We propose that one unbalance, or several, caused by the deletion of the BTN2 gene triggers a transcriptional response. We have carried out a global transcriptomic analysis in the btn2 mutant, and found many genes whose transcription is altered in the mutant, but that are not significantly involved in cell-to-cell interaction or stress response (data not shown). It is tempting to speculate that Btn2p affects the right localization of flocculins in the cell wall in the same way that it affects the localization of the plasma membrane channel Ist2p. This may cause an unbalance that activates the FLO11 transcription and results in better mat formation. The BTN2 deletion affects adhesion to plastic; therefore, it may decrease the expression or localization of one additional protein, or some, that are relevant to this phenotype.
As far as we are aware, this is the first time that an industrial flor yeast has been genetically manipulated. Previous studies have been carried out on haploid spores derived from flor yeasts (Zara et al., 2005; Fidalgo et al., 2006). In this work, we have not encountered any particular difficulty compared with the other industrial wine yeasts that we have modified previously (Cardona et al., 2007). Therefore, this opens a way to study and improve this kind of yeast strains. The manipulation of the BTN2 gene has led to a complex phenotype caused by its pleiotrophic roles, but it has indicated that Btn2p is a very attractive protein that links stress resistance to biofilm formation. Thus, it is worth conducting an in-depth study into this protein to gain a better understanding of these processes.
This work has been supported by grants AGL2005-00508 from the ‘MEC’ (Ministry of Education and Science) and GRUPOS03/012 and GVACOMP2007-157 from the ‘Generalitat Valenciana’ to E.M. and A.A., and by grants PETRI 95-0855 OP from DGICYT to J.M.C.