AQY1 and AQY2 were sequenced from five commercial and five native wine yeasts. Of these, two AQY1 alleles from UCD 522 and UCD 932 were identified that encoded three or four amino-acid changes, respectively, compared with the Σ1278b sequence. Oocytes expressing these AQY1 alleles individually exhibited increased water permeability vs. water-injected oocytes, whereas oocytes expressing the AQY2 allele from UCD 932 did not show an increase, as expected, owing to an 11 bp deletion. Wine strains lacking Aqy1p did not show a decrease in spore fitness or enological aptitude under stressful conditions, limited nitrogen, or increased temperature. The exact role of aquaporins in wine yeasts remains unclear.
Forming one of the largest subgroups in the major intrinsic protein (MIP) family, aquaporins have been identified in many different organisms (Zardoya, 2005). MIPs form channels across biological membranes that control recruitment of water and small solutes such as glycerol and urea in all living organisms (Benga, 2003). MIP-related proteins are characterized by a topology of six transmembrane spanning domains and share a highly conserved NPA (Asn-Pro-Ala) motif in each of the two most prominent loops (Borgnia et al., 1999). The suggestion that these NPA segments line the path of water permeation, and the fact that the two halves of the protein are homologous, has led to the proposed hourglass structural model, in which the NPA loops overlap in the middle of the membrane creating a narrow aqueous channel (Jung et al., 1994). The oligomerization of aquaporins in the yeast Saccharomyces cerevisiae is not understood in detail, but studies with the human erythrocyte aquaporin AQP1 have indicated that the protein is a tetramer composed of functionally independent aqueous pores (Walz et al., 1997).
Two aquaporin encoding genes (AQY1 and AQY2) have been identified in the yeast genome and, surprisingly, different laboratory and wild-type strains of S. cerevisiae were found to have specific sequential and functional differences in these genes. Most laboratory strains, including the S. cerevisiae database strain (S288C), have been shown to contain nonfunctional alleles of both aquaporin genes (Laizé, 2000). However, the somewhat unusual strain Σ1278b, which is also used to study pseudohyphal development, appears to have functional alleles of both aquaporin genes. While the water transport function of Aqy1pΣ1278b has been demonstrated in the heterologous oocyte system (Bonhivers et al., 1998), water transport via Aqy2pΣ1278b was not observed (Carbrey et al., 2001). The water channel activity of Aqy2pΣ1278b was later demonstrated by stopped-flow analysis of microsomal vesicles (Meyrial et al., 2001).
The physiological role of yeast aquaporins is not well understood, but gene and protein expression data suggest that the two 83% identical gene products may have different roles. AQY1 mRNA levels have been observed to be higher in strain SK1 during meiosis and sporulation (Chu et al., 1998), during nitrogen limitation in UCD 2100 (Backhus et al., 2001), and under heat stress during fermentation of synthetic grape juice with strain UCD 713 (Mangahas and Bisson, unpublished results). These data suggest a role for AQY1 in stress response. Interestingly, Meyrial (2001) have reported that the AQY1 gene product could not be detected in vegetative cells of Σ1278b, but appears to be expressed abundantly in yeast spores where Aqy1p may be involved in spore maturation and/or germination. In fact, Sidoux-Walter (2004) report that Aqy1p is involved in spore formation in the SK1 background, and that spore fitness is reduced in the absence of Aqy1p. AQY2, on the other hand, is expressed in proliferating Σ1278b cells, and the mRNA diminishes after nutrient depletion (Meyrial et al., 2001). In addition, expression of AQY2 (in the Σ1278b background) may be activated by the protein kinase A catalytic subunit Tpk2p, and is diminished in a HOG-dependent manner (Robertson et al., 2000). These observations could indicate that Aqy2p plays a role in water efflux in turgor control during rapid growth and under low-osmolarity conditions. Tanghe (2004) have shown that Aqy2p also has a role in freeze tolerance. These observations beg the question of whether AQY1 and/or AQY2 are expressed in wine yeasts and, if functional, what role these proteins play in stress response during fermentation.
Laizé (2000) have shown that several natural isolates and industrial yeast strains contain functional alleles for AQY1, yet have nonfunctional alleles of AQY2. They concluded that natural and industrial conditions provide selective pressure to maintain AQY1. Of the 52 strains surveyed, only three were identified as wine yeasts, and they were either haploid or polyploid. In addition, they used restriction digest patterns to identify different AQY1 and AQY2 alleles, and, therefore, could not account for any point mutations that might affect the function of the protein in strains exhibiting the wild-type restriction pattern for AQY1. We have extended this research by surveying the aquaporin genes in ten diploid wine yeast strains by DNA sequencing and verifying the water channel activity of representative alleles in the Xenopus expression system. To define the role of Aqy1p in wine strain fermentations, AQY1 was deleted in the UCD 932 strain, which showed the highest activity in the Xenopus oocyte analysis. As the physiological role of yeast aquaporins is still not well understood, this research aims to identify the role aquaporins play in native yeast isolates under conditions mimicking their natural environment.
Strains and culture conditions
Saccharomyces cerevisiae strains utilized were five native wine strains (UCD 932, 935, 937, 939, and 940) and five commercial wine strains (UCD 522, 713, 905, 2031, and 2100), all of which are stable, diploid yeasts. The former represent yeasts isolated from Italian wineries and cultured in the laboratory (Mortimer et al., 1994), while the latter are commercial yeasts obtained from the UCD Yeast Culture Collection. Three other yeasts were used for comparison: Saccharomyces chevalieri, a species found to have a functional AQY2 gene (Carbrey et al., 2001), SK1, a fast sporulating, S. cerevisiae strain used in meiotic studies (Chu et al., 1998), and X2180, a strain genetically identical to the S. cerevisiae database strain, S288C (Mortimer & Johnston, 1986). Cultures were grown in YEPD liquid medium (1% yeast extract, 2% bactopeptone, and 2% dextrose) by incubation for 12 h at 25°C with agitation on a roller drum. Spores were prepared by transferring these cultures into 0.5% potassium acetate liquid medium and incubating at 25°C for 3 days (Sidoux-Walter et al., 2004). Spores were then separated and allowed to germinate on YEPD solid medium. Escherichia coli DH5α were used for cloning experiments. Escherichia coli cultures with plasmid constructs were prepared in Luria–Bertani (LB) liquid medium (0.5% yeast extract, 1% bactotryptone, and 0.5% NaCl), supplemented with 100 μg mL−1 ampicillin, by incubation for 12 h at 37°C with agitation.
Sequencing of the AQY1 and AQY2 genes was performed according to standard procedures. Genomic DNA was collected from each strain and used as a template for PCR reactions. A typical PCR reaction consisted of 30 cycles: 94°C for 1 min, 52°C for 1 min, and 68°C for 1 min, using Platinum Taq HIFI DNA Polymerase (Invitrogen, Carlsbad, CA). AQY1 was amplified using primers JKAQY1F1 (5′-TGGTGCTGTC TGTCAATACG-3′) and JKAQY1R5 (5′-GTCAAGGTCGGC TATAAAG-3′). AQY2 was amplified using primers JKAQY2F1 (5′-TTCTTGTTCCTGGCTATT-3′) and JKAQY2R4 (5′-ACGATGGGAGCGTTATGC-3′). Alignment studies were performed using BioEdit© software. The nucleotide sequences for AQY1 from UCD 522 and 932 were submitted to the GenBank™ Databank with accession numbers DQ338534 and DQ338535, respectively.
AQY1 was deleted from wine strains by the one-step gene replacement method using the geneticin resistance cassette loxP-KanMX-loxP (amplified from plasmid pUG6) (Gueldener et al., 2002). After transformation, cells were incubated for 2–4 h at 25°C in YEPD, washed once in sterile water, and plated on YEPD plates (2.5% agar) plus 200 mg L−1 G418 (Sigma, St. Louis, MO). After 48 h at 30°C, the colonies were restreaked on fresh YEPD plates plus G418 to confirm resistance. Homozygous aqy1-null mutants (1B and 1C) were generated by sporulation of heterozygous AQY1/aqy1 mutants and separation of the resulting spores that self-diploidize (Parental = 1A and 1D). To determine the neutrality of the dominant geneticin resistance gene (KanMX) during fermentation, the double ho-null mutant was generated as above (HO is a gene involved in mating-type switching). The ho-null strain was compared with the parental strain to see if expression of the KanMX gene affected the fermentation profile. All gene knockouts were verified by PCR and Southern blotting (data not shown).
Plasmid pAQP1 was kindly provided by Dr Katherine Wasson (UC Davis). Xenopus expression plasmid pUMA23 [containing an SP6 RNA polymerase recognition site and 5′-untranslated region (5′-UTR) of the Xenopusβ-globin gene] was obtained from American Type Culture Collection (#87303). For construction of the pUMAQP1 plasmid, the AQP1 coding region and 3′-UTR were amplified by PCR from pAQP1 with JKAQP1F1 (5′-TGGCAGATGAATTCCGCAC-3′) and JKAQP1R1 (5′-CGGTGGCGGCCGCTCTAG-3′) as an EcoRI–NotI fragment. This fragment was ligated into the EcoRI–NotI site of the pUMA23 vector resulting in the human AQP1 coding region flanked by the 5′-UTR and 3′-UTR of the Xenopusβ-globin gene. For construction of the pUMAQY plasmids, representative alleles of AQY1 were amplified as NcoI–BstEII fragments by PCR from genomic DNA using NcoIF1 (5′-TAACTATACCATGGCTTCG-3′) and BstE2R1 (5′-TAAAGGTTACCGTCAAGGTCGGCTATAAAG-3′), and representative alleles of AQY2 were amplified using NcoIF2 (5′-AGTTATACCATGGCTAACG-3′) and BstE2R2 (5′-TAAAGGTTACCACGATGGGAGCGTTATGC-3′). These fragments were ligated into the excised NcoI–BstEII site of pUMAQP1 resulting in the AQY1 or AQY2 coding region flanked by the 5′-UTR and 3′-UTR of the Xenopusβ-globin gene. In creating the NcoI site in primers NcoIF1 and NcoIF2, a T-to-G substitution (in bold) at the fourth base in the coding sequence resulted in an S-to-A amino-acid change. As this change is close to the N-terminus of the protein, we do not think this will have any functional effect on the proteins. Plasmid constructions were verified by restriction and sequence analysis.
Xenopus oocyte expression
Representative AQY1 and AQY2 alleles were cloned into the pUMA23 expression vector and capped cRNA was produced using SP6 RNA polymerase (Promega, Madison, WI). Approximately 10 ng cRNA (50 nL) was micro-injected into Stage V/VI oocytes taken from female Xenopus laevis. Water channel activity was confirmed by transferring oocytes from high (200 mosM) to low (70 mosM) osmolar Modified Barth's solution [88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO3, 15 mM Tris-HCl (pH 7.6), 0.3 mM Ca(NO3)2, 0.4 mM CaCl2, 0.8 mM MgSO4, 0.1 mg mL−1 Penicillin-G, and 0.1 mg mL−1 streptomycin sulfate] after 72 h incubation at 16°C and observing a rapid (≈2 min) increase in oocyte volume, with water-injected oocytes used as a negative control. Swelling assays were performed at 25°C and digital images of oocytes were captured at 4 s intervals. The surface area of images was calculated using MetaMorph 6.2 software (Molecular Devices, Downingtown, PA). Relative volume of oocytes and osmotic water permeability (Pf) were calculated as described previously (Preston et al., 1993). Human AQP1 cRNA from the pUMAQP1 plasmid was used as a positive control.
Fermentations were performed in triplicate in synthetic minimal must medium (MMM) containing 11% or 13% of both glucose and fructose, for a total of 22% or 26% sugar, at 25 or 35°C with agitation (Spiropoulos et al., 2000). Assimilable nitrogen levels of 131, 216, or 467 mg L−1 were made by varying the amounts of arginine and diammonium phosphate added to the medium. Fermentations were carried out in 300 mL of medium contained in 500 mL Erlenmeyer flasks, closed with foil to achieve semi-anaerobic conditions. Strains were grown to stationary phase in MMM and then used to inoculate 300 mL medium with a starting OD600 nm≈0.03. Cell growth and ethanol production were monitored by optical density and flask weight loss, respectively, until flask weight remained constant, after approximately 7–10 days. Residual sugar was determined using the Clinitest™ reducing sugar test (Fischer Scientific, Pittsburg, PA). Clinitest™ is a standardized method for the quantitative determination of reducing sugars by copper reduction, which results in a color change that varies with the amount of reducing substances present.
Results and discussion
Polymorphism of aquaporin genes in wine yeasts
The AQY1 and AQY2 genes were sequenced from five native and five commercial wine yeast strains. The Aqy1p sequence was found to be mostly identical to all the wine strain sequences and Σ1278b, except that three substitutions were found in all the wine strain sequences: R42K, V53A, and P308S (Fig. 1). Bonhivers (1998) have shown that strain GRF5 (Saccharomyces norbensis) contains the exact same three substitutions; however, they did not report on whether this allele showed water channel activity in the Xenopus expression system. One variant sequence, which contained an additional G226S substitution, was identified in strain UCD 932 (Fig. 1). This amino-acid substitution may have a functional effect on Aqy1p because of its proximity to the second NPA region (residues 230–232). AQY1 sequenced from strains SK1 and Saccharomyces chevalieri was identical to the Σ1278b sequence (data not shown). AQY1 sequenced from strain X2180 was identical to strain S288C (data not shown), containing the debilitating N47D, V121M, and P255T mutations as described previously (Fig. 1) (Bonhivers et al., 1998).
AQY2 in all 10 wine strains contains an 11 bp deletion that fragments the gene into two open reading frames (data not shown). This interruption is identical to the X2180 and S288C strain sequences (Carbrey et al., 2001; data not shown), and, therefore, AQY2 is also thought to be nonfunctional in the wine strains. Aqy2p from S. chevalieri was identical to Σ1278b, save for a P141S substitution. This was the only Saccharomyces species previously found to have a functional Aqy2p when expressed in oocytes (Carbrey et al., 2001). AQY2 from laboratory strain SK1 was found to contain a single nucleotide deletion (+25), much like other laboratory strains sequenced by Laizé (2000), which apparently leads to a truncated and nonfunctional protein. These results compare favorably with the results obtained by Laizé (2000), which concluded that the three wine strains might contain functional alleles of AQY1, but did not have functional alleles of AQY2, based upon restriction analysis. No differences were found between both alleles of AQY1 or AQY2 in each diploid wine strain.
Oocyte swelling assay with wine yeast aquaporins
Osmotic water permeability (Pf=mean±SD) of Xenopus oocytes was measured after microinjection with 50 nL of water or 50 nL of water containing approximately 10 ng of aquaporin cRNA. After incubation for 72 h, the oocytes were transferred from 200 to 70 mosM Modified Barth's Solution, and swelling was measured. Water-injected oocytes swelled minimally (Pf=13±6, n=18), whereas all human AQP1 RNA-injected oocytes swelled rapidly (Pf=104±23, n=8) (Fig. 2). There was a sixfold increase (Pf=83±29, n=19) in Pf in oocytes injected with SK1 AQY1 RNA, which is identical to the sequence of Σ1278b AQY1. This value is smaller than the value for Σ1278b AQY1 (Pf≈125) published in Bonhivers (1998), but this may be due to experimental differences. Oocytes expressing Aqy1p from UCD 522 showed a threefold increase in Pf over water-injected oocytes (Pf=34±12, n=15), whereas those expressing Aqy1p from UCD 932 had much more activity (Pf=81±23, n=18).
The G226S substitution in UCD 932 apparently overcomes the otherwise detrimental substitutions in the other wine strain AQY1 alleles. Crystallization data for S. cerevisiae Aqy1p have not been published to date; however, the location of the four amino-acid changes in UCD 932 Aqy1p can be compared with the published crystallization data for human AQP1. The R42K and V53A substitutions, while proximate to the first transmembrane region of Aqy1p (Fig. 1), are conservative, and therefore probably do not have much of a functional effect on water transport. The P308S substitution results in a major structural change, as proline residues introduce kinks in protein structure. The C-terminal tail of Aqy1p has been implicated in the regulation of water transport (Bonhivers et al., 1998), so any change in the structure or spatial movement of the tail likely has a functional effect on water transport. The G226S substitution in UCD 932 is located very near the second NPA box (Fig. 1) and may positively affect water transport. This position is comparable to the highly conserved threonine 187 in human AQP1 (Heymann et al., 1998). T187 is located on the extracellular side of the aquaporin and in the mouth of the channel close to the pathway of water transport (Heymann et al., 1998). There is a possibility that the substituted serine participates in channel stability and/or recruits/stabilizes water molecules with its hydrogen-bonding capability and increases the rate of water transport.
Oocytes expressing Aqy2p from all wine strains and laboratory strain SK1 failed to show a significant increase in Pf (Fig. 2), while oocytes expressing S. chevalieri Aqy2p (Pf=75±12, n=6) exhibited a sixfold increase over water-injected oocytes. This was consistent with the Pf value (≈100) attributed to S. chevalieri Aqy2p in Carbrey (2001). While these calculated values indicate that the minimal substitutions found in the wine strain AQY1 sequences do have an effect on the water channel activity of Aqy1p in the Xenopus expression system, we cannot conclude that these aquaporins are functional in yeast. Yet, AQY2 in wine yeasts is almost certainly nonfunctional.
Effect of AQY1 disruption on wine yeast fermentation and sporulation
As the aquaporin from strain UCD932 showed the highest water channel activity of the wine strains, it was selected for further analysis of the role of these genes in wine yeasts. The AQY1 gene was deleted in UCD 932 and the single knockout strain was sporulated to produce a double aqy1-null strain. Sidoux-Walter (2004) reported a 20–50% decrease in spore viability (germination) in an aqy1-null SK1 strain and concluded that Aqy1p has a functional role in spore formation. In order to see whether Aqy1p plays a role in sporulation in UCD 932, the double knockout was sporulated and the dissected spores were allowed to germinate. This strain is homothallic, and thus diploidizes following sporulation. A heterozygous diploid (AQY1/aqy1-null) was sporulated and the four spores allowed to diploidize. The four homozygous diploids were then allowed to sporulate. Six complete tetrads from each homozygous diploid were obtained and the tetrads dissected. Spore viability in the double knockout diploids was identical to that in the UCD 932 parental strain (48 out of 48 spores germinated). This suggests either that UCD 932 can overcome the deficiency in Aqy1p in other ways to produce viable spores, or that this gene is not truly necessary for spore fitness in wild backgrounds. This is also true for many laboratory strains that do not possess functional Aqy1p but show high spore viability (Sidoux-Walter et al., 2004).
The wine yeast strain UCD 932 was also used to assess the role of AQY1 during small-scale fermentations of synthetic grape juice. The double knockout was utilized in side-by-side fermentations with the parental strain under varying conditions to compare the enological aptitude of both strains. Previous gene expression experiments have shown an increase in AQY1 mRNA during nitrogen limitation or heat stress in small-scale fermentations (Backhus et al., 2001; unpublished results), and as the protein may have a role in stress response, the absence of Aqy1p in wine yeasts undergoing stress (e.g. nitrogen limitation and increased temperature and ethanol levels) may have a detrimental effect on the fermentation process.
The fermentation profile of the parental strain was compared with both the heterozygous (AQY1/aqy1) and homozygous aqy1-null strains under differing conditions. As the aqy1-null strain was produced by sporulating the heterozygous strain, the aqy1-null strains (1B and 1C) may not be genetically identical to the parental strain. To determine whether any genetic variation of the resulting spores had an effect on the enological aptitude, we also tested the ‘wild-type’ strains (1A and 1D) in side-by-side fermentations with UCD 932. There were no significant differences in the fermentation profile of strains 1A and 1D vs. UCD 932 under every condition tested (data not shown). Therefore, sporulation did not have a significant effect on the enological aptitude of the daughter spores.
Strains 1B, 1C, and the heterozygous strain (AQY1/aqy1Δ) were tested in fermentations vs. UCD 932 to determine the effect of nitrogen limitation on enological aptitude. Initial assimilable nitrogen levels were varied between 467 (high) and 131 mg L−1 (low) nitrogen with 11% glucose and 11% fructose at 25°C. All of the fermentations proceeded in a typical fashion (data not shown), lasting about 225 h until reaching dryness (defined as <0.5% sugar remaining), and there were no significant fermentation differences between the aqy1-null strain, the heterozygous strain, and UCD 932 (data not shown). There were also no apparent growth defects between the strains during fermentation as monitored by optical density (data not shown).
Strains 1B and 1C were also tested under conditions of heat and ethanol stress. Comparing the fermentations between strains 1B, 1C, and UCD 932 at 35°C and with 11% glucose and 11% fructose, the fermentations started at a quicker pace but the cultures did not attain the maximal cell concentration, and none achieved dryness (data not shown). There was some variability in the fermentation profile between strains at 35°C, but no obvious pattern was observed for the aqy1-null strains (data not shown). Comparing the strains in a high sugar (13% glucose and 13% fructose) fermentation at 25°C, which would result in an increased final ethanol concentration, all strains reached dryness and there was little difference in fermentation profile (data not shown).
To determine the neutrality of the geneticin resistance gene KanMX during fermentation, the ho-null strain was compared with UCD 932 under all conditions. No significant differences were found in the fermentation profile of the ho-null strain and the profile of the parental strain (data not shown), indicating that expression of the KanMX gene does not have a significant effect on the fermentation profile.
The wine yeast strains investigated in this study are not dependent on aquaporins to adapt to stress under certain fermentation conditions. These yeasts contain a functional AQY1, but not a functional AQY2. Different alleles for AQY1 were found in the wine strains that differed in their capacity to transport water. The G226S mutation in the UCD932 AQY1 increases the osmotic water permeability in the Xenopus expression system. These findings support the concept that the natural environment selects for a functional AQY1 in wild yeasts (Laizé, 2000). Yet, wine yeasts lacking AQY1 do not show a decrease in either spore fitness or in their ability to ferment synthetic grape juice, suggesting that wine yeasts have alternate methods to adapt to stress and produce viable spores, if necessary. Further, this analysis underscores the difficulty in using expression studies to define function. The expression of AQY1 under stress conditions in UCD932 does not necessarily mean this gene is involved in or required for the stress response (Backhus et al., 2001). Perhaps water transport through other routes is sufficient to adapt to a changing environment, or passive diffusion through the plasma membrane may also be sufficient. Likely, we have not discovered the very specific conditions under which aquaporins offer an advantage to the cell during adaptation. The exact function of AQY1 during wine fermentation is an interesting question for future studies. Varela (2005) recently concluded that the genome of wine strains contains a significant number of genes not found in the sequence of S288C, and there may be more functional redundancy in native and commercial isolates than in laboratory strains.
We are very grateful for the support received from Dr Jie Zheng and colleagues (Department of Physiology and Membrane Biology, UC Davis). Dr Zheng generously donated oocytes and microscope time for the completion of Xenopus expression experiments. We would also like to thank Dr Jennifer Carbrey (Johns Hopkins) for the gift of the Saccharomyces chevalieri strain, Joshua Chang-Mell for the gift of the SK1 laboratory strain, and Dr An Tanghe (Catholic University of Leuven, Belgium) for kindly providing the plasmid pUG6. This research was supported by an NSF Graduate Student Research Fellowship and a grant from the American Vineyard Foundation and the California Competitive Grant Program for Research in Viticulture and Enology.