|Home | About | Journals | Submit | Contact Us | Français|
Successful fermentations to produce ethanol require microbial strains that have a high tolerance to glucose and ethanol. Enhanced glucose/ethanol tolerance of the laboratory yeast Saccharomyces cerevisiae strain BY4741 under certain growth conditions as a consequence of the expression of a dominant mutant allele of the SPT15 gene (SPT15-300) corresponding to the three amino acid changes F177S, Y195H, and K218R has been reported (H. Alper, J. Moxley, E. Nevoigt, G. R. Fink, and G. Stephanopoulos, Science 314:1565-1568, 2006). The SPT15 gene codes for the TATA-binding protein. This finding prompted us to examine the effect of expression of the SPT15-300 allele in various yeast species of industrial importance. Expression of SPT15-300 in leucine-prototrophic strains of S. cerevisiae, Saccharomyces bayanus, or Saccharomyces pastorianus (lager brewing yeast), however, did not improve tolerance to ethanol on complex rich medium (yeast extract-peptone-dextrose). The enhanced growth of the laboratory yeast strain BY4741 expressing the SPT15-300 mutant allele was seen only on defined media with low concentrations of leucine, indicating that the apparent improved growth in the presence of ethanol was indeed associated with enhanced uptake and/or utilization of leucine. Reexamination of the microarray data published by Alper and coworkers likewise suggested that expression of genes coding for the leucine permeases, Tat1p and Bap3p, were upregulated in the SPT15-300 mutant, as was expression of the genes ARO10, ADH3, ADH5, and SFA1, involved in leucine degradation.
Improvement of stress tolerance in microorganisms applied in industrial fermentations for the production of ethanol is of major interest (26, 34). Based on screens for ethanol sensitivity/tolerance in Saccharomyces cerevisiae (12, 16-18, 35, 37, 40), it appears that this trait in yeast is possibly controlled by several genes acting in concert. Using global transcription machinery engineering (gTME), a tool to reprogram gene transcription for eliciting new phenotypes important for technological applications, Alper et al. (2) found mutants of S. cerevisiae with improved glucose/ethanol tolerance. In that work, mutated versions of the SPT15 gene, which codes for the TATA-binding protein, were generated by random in vitro mutagenesis and expressed in the laboratory strain BY4741. The authors identified one dominant allele, SPT15-300, which corresponds to the three amino acid changes F177S, Y195H, and K218R, that conferred increased tolerance of the yeast to ethanol (2). Although extensive analyses, such as transcriptional profiling and deleting and overexpressing individual genes, were carried out, a particular pathway or a genetic network responsible for the observed growth gain of the SPT15-300-expressing strain could not be identified (2). During our attempts to analyze the effect of the mutant SPT15-300 alleles in various yeast species of industrial importance, we discovered that the described improvement of growth in the presence of ethanol of the standard laboratory strain BY4741, the strain used by Alper et al. (2), is associated with improved uptake and/or utilization of leucine on media containing small amounts of leucine.
The Saccharomyces strains investigated in this study were S. cerevisiae strains BY4741 (MATa his3ΔD1 leu2Δ0 met15Δ0 ura3Δ0) (5), in which the LEU2 gene is completely deleted (obtained from Euroscarf, Frankfurt, Germany), and Y55 (23) derivative JT20150 (MATα MAL1) (obtained from J. M. Thevelein, Katholieke Universiteit, Leuven, Belgium); S. bayanus NRRL Y-11845 (MCYC 623) (7, 20) (provided by C. P. Kurtzman, Microbial Genomics and Bioprocessing Research Unit, Peoria, IL); and S. pastorianus W-34/70 (25) (obtained from Hefebank Weihenstephan, Freising, Germany).
Yeast cells were cultured aerobically in complex rich medium YPD (1% [wt/vol] yeast extract, 2% [wt/vol] peptone), 2% [wt/vol] glucose) (32) supplemented when necessary with G418 (final concentration of 100 or 300 μg/ml as indicated), synthetic complete minimal (SC) medium (6.7 g/liter yeast nitrogen base [without amino acids] supplemented with amino acids as specified in reference 32) without uracil (SC−Ura) and a modified composition containing five times the amount of leucine (i.e., 150 mg/liter instead of 30 mg/liter) (SC−Ura 5 × Leu), SC lacking leucine (SC−Leu), yeast synthetic complete (YSC) medium (6.7 g/liter yeast nitrogen base [without amino acids] supplemented with Qbiogene CSM-URA [a commercial amino acid mixture]) lacking uracil (containing 100 mg/liter leucine) as described by Alper et al. (2), and YSC lacking leucine (prepared as described for YSC−Ura, with Qbiogene CSM-LEU ). SC media were buffered (pH 5.5) with 1% (wt/vol) succinic acid and 0.6% (wt/vol) NaOH. SC and YSC media were supplemented with glucose and/or ethanol as indicated. S. cerevisiae strains were incubated at 20 or 30°C (as indicated), and S. bayanus and S. pastorianus were cultivated at 20°C. Saccharomyces species were transformed by use of the lithium acetate method (3).
Escherichia coli strain DH5α (Invitrogen A/S, Taastrup, Denmark) was used for plasmid selection/propagation and cultivated as described previously (31).
Standard recombinant DNA manipulations were performed as described previously (31). DNA-modifying enzymes were obtained from Invitrogen (Invitrogen A/S, Taastrup, Denmark), New England Biolabs (Medinova Scientific A/S, Glostrup, Denmark), and Promega (Promega Biotech AB, Nacka, Sweden) and used as recommended by the suppliers. PCRs were carried out with Phusion high-fidelity DNA polymerase (Finnzymes, Medinova Scientific A/S, Glostrup, Denmark). DNA sequencing and oligonucleotide synthesis were performed by Eurofins MWG (Ebersberg, Germany); oligonucleotide sequences are available on request.
(i) One set of SPT15 variants (see below) was inserted into vector pCJR2 (a CEN-based vector with a native S. cerevisiae TEF1 promoter and G418 selection), which was constructed by cloning a 852-bp SacI-PvuII-fragment of p416TEF (24) (obtained from ATCC, LGC Standards AB, Boras, Sweden) containing the wild-type S. cerevisiae TEF1 promoter into a 4471-bp SacI-EcoRV-digested vector fragment of pCJR1. This plasmid was constructed by cloning a 1,447-bp BglII (blunt ended by DNA polymerase [Klenow fragment])-SacI fragment (KanMX4 cassette) of pUG6 (15) (obtained from Euroscarf, Frankfurt, Germany) into a 3,082-bp TthIII1 (blunt ended with Klenow fragment)-SacI vector fragment of pRS315 (33).
(ii) The second set of SPT15 expression vectors was constructed identically to pCJR2, except that the wild-type S. cerevisiae TEF1 promoter was exchanged with mutant version 2 as described previously (1, 27). To accomplish this, a 403-bp SacI-XbaI-digested synthetic DNA fragment (GenScript, Piscataway, NJ, USA) containing the mutant TEF1 promoter (1) was used to replace the native SacI-SpeI-digested promoter.
(iii) A third set of SPT15 vectors (CEN-based vector, mutant S. cerevisiae TEF1 promoter, URA3 selection) was constructed by inserting SacI-EagI-digested fragments of pCJR7 (1,434-bp fragment with S. cerevisiae-type SPT15) or pCJR8 (1,430-bp fragment with S. cerevisiae-type SPT15-300) into a 4,805-bp SacI-EagI-digested vector fragment of p416TEF.
(iv) A fourth set of SPT15 vectors (CEN-based vector, mutant S. cerevisiae TEF1 promoter, LEU2 selection) was constructed by inserting SacI-EagI-digested fragments of pCJR7 (1,434-bp fragment with S. cerevisiae-type SPT15) or pCJR8 (1,430-bp fragment with S. cerevisiae-type SPT15-300) into a 6,005-bp SacI-EagI-digested vector fragment of pRS315.
The lager brewing yeast, Saccharomyces pastorianus, is a hybrid of S. cerevisiae and a Saccharomyces species related to S. bayanus (21, 25). Genes in the genome of lager brewing yeast that have high identity with genes found in S. cerevisiae are called S. cerevisiae type, while genes more distantly related are called non-S. cerevisiae type. SPT15 gene variants (Table (Table1)1) were obtained as follows: (i) a 753-bp BamHI (blunt ended with Klenow fragment)-SpeI-digested S. cerevisiae-type SPT15 fragment was amplified from S. cerevisiae BY4741 genomic DNA by PCR, (ii) a 749-bp SpeI-SmaI-digested S. cerevisiae-type SPT15-300 fragment was obtained as a synthetic gene (GenScript, Piscataway, NJ) according to the mutant sequence as described by Alper et al. (2), (iii) a 775-bp SpeI-EcoRV-digested non-S. cerevisiae-type SPT15 fragment was amplified from S. pastorianus W-34/70 genomic DNA by PCR, and (iv) a fragment containing non-S. cerevisiae-type SPT15-300 was constructed by cloning a 259-bp SpeI-BglII-digested 5′ fragment of non-S. cerevisiae-type SPT15 (in which the BglII site was introduced by silent mutation using PCR) to a 504-bp BglII-SmaI digested 3′ fragment of S. cerevisiae-type SPT15-300 (exchange of gene fragments was possible, since the encoded S. cerevisiae-type and non-S. cerevisiae-type Spt15 proteins differ only at amino acids 31 and 36). The correct sequence of each vector was confirmed by DNA sequencing.
The growth phenotypes of SPT15 transformants were examined as described by Alper et al. (2). In short, yeast transformants were precultured in YSC (2) or SC (32) medium containing various amounts of glucose as indicated and diluted to an optical density at 600 nm (OD600) of 0.01 in fresh medium supplemented with various amounts of ethanol as indicated. The OD600 was measured after 20 h of incubation at 30°C with shaking. In the case of G418 selection, YSC media were supplemented with 300 μg/ml G418. The ethanol tolerance of SPT15 transformants was analyzed on plate assays in which solid medium (as indicated) was supplemented with 6% or 8% ethanol (as indicated). Tenfold serial dilutions of cell cultures, pregrown in appropriate media (as indicated), at an OD600 of 1.0 (initial dilution) were spotted on plates. Growth assays were performed in triplicate. Results of representative experiments are shown.
The microarray data (accession no. GSE5185) were downloaded from the Geo Expression Omnibus database (4) and analyzed using R and Bioconductor (13). Array “GSM116825” (SPT15-300 plus 60 g/liter glucose and 5% ethanol) was identified as an outlier and removed, and only probes specific for S. cerevisiae were used in our analyses. rma was used for quantile normalization and probe index calculations, and these were subsequently normalized using Qspline (19, 39). For statistical testing, two-factor analysis of variance was used, with the factors “genotype” (i.e., wild type versus SPT15-300) and “medium” (i.e., 20 g/liter glucose [medium A] versus 60 g/liter glucose and 5% ethanol [medium B]). The false-discovery rate (FDR) was estimated using a Monte Carlo approach, and statistical significance was set at an FDR of 0.005.
Encouraged by the report by Alper and coworkers (2), we were interested in applying gTME to yeasts in order to improve their ethanol tolerance and ultimately fermentation performance. As a first step, we decided to evaluate the effect of the SPT15-300 mutant allele identified by Alper et al. (2) in various yeast species of industrial importance. The lager brewing yeast, Saccharomyces pastorianus, is a hybrid between S. cerevisiae and a Saccharomyces species related to S. bayanus (21, 25). Genes in the genome of lager brewing yeast that show a high percentage of sequence identity to genes found in S. cerevisiae are called S. cerevisiae type, while genes more distantly related are called non-S. cerevisiae type. We introduced the three point mutations identified in SPT15-300 in both types of genes, and the wild-type and mutant SPT15 genes were subsequently inserted into plasmids under the control of the wild-type S. cerevisiae TEF1 promoter. Since ethanol sensitivity/tolerance screens are generally performed in rich complex media, i.e., YPD supplemented with various amounts of ethanol ranging from 6 to 12.5% (12, 16-18, 35, 37, 40), we analyzed the growth of SPT15 transformants of S. cerevisiae BY4741 and Y55 (JT20150), S. bayanus NRRL Y-11845, and S. pastorianus W-34/70 on rich complex solid medium (i.e., YPD) supplemented with 8% ethanol. This ethanol percentage was arbitrarily chosen in order to analyze “ethanol-resistant” and “ethanol-sensitive” yeasts (such as S. cerevisiae Y55 and S. pastorianus W34/70, respectively) under one single condition. Unfortunately, none of the yeasts harboring the mutant SPT15-300 gene displayed the expected improved ethanol tolerance (Fig. (Fig.1).1). As a control and in order to repeat the experiments described by Alper et al. (2), plasmids that carried the wild-type and mutant SPT15 genes under the control of the weaker mutant version of the S. cerevisiae TEF1 promoter were constructed (1, 2, 27). In agreement with the findings of Alper et al., we found that the S. cerevisiae laboratory strain BY4741 transformed with SPT15-300 showed an apparent increase in ethanol tolerance in liquid YSC medium (2) regardless of whether the SPT15 genes were expressed from a URA3- or a G418 selection-based plasmid (Table (Table2).2). Evaluation of the contribution of promoter strength to the appearance of ethanol tolerance in cells expressing SPT15-300 (i.e., comparison of SPT15 expression using the native and mutant S. cerevisiae TEF1 promoters) demonstrated that the effect was stronger when the native promoter was used than when the weaker mutant version was used (data not shown). The enhanced ethanol tolerance of cells carrying the SPT15-300 allele was also apparent on SC plates with 30 mg/liter leucine but not on SC plates containing 150 mg/liter leucine (Fig. (Fig.2A).2A). When transformants were grown in YSC medium that contained 100 mg/liter leucine, prepared as described by Alper et al. (2), the enhanced ethanol tolerance was only marginally manifested (Fig. (Fig.2A).2A). In the absence of ethanol, the enhanced growth of cells with the SPT15-300 allele was also noticeable on SC−Ura medium with 30 mg/liter leucine (Fig. (Fig.2A).2A). Thus, the apparent improved ethanol tolerance could be related to the improved growth of the SPT15-300 mutant in media containing smaller amounts of leucine. The S. cerevisiae laboratory strain BY4741 is deficient in leucine biosynthesis due to the deletion of the LEU2 gene, encoding β-isopropylmalate dehydrogenase, the third enzyme in leucine biosynthesis (5). Therefore, we examined the effect of the SPT15 wild-type and mutant alleles inserted into a LEU2-containing plasmid in BY4741 cells transformed to leucine prototrophy. Cells containing either the wild-type or the mutant allele grew equally well on solid media lacking leucine (SC−Leu) whether or not ethanol was present (Fig. (Fig.2B).2B). This indicated that the observed increased growth of cells harboring the SPT15-300 allele indeed is related to improved uptake and/or utilization of leucine.
We also tested the growth of the transformed cells in liquid SC-based media containing different amounts of glucose (Fig. (Fig.3).3). In cells transformed with URA3-based plasmids, the SPT15-300 mutant showed enhanced growth in SC−Ura medium containing 30 mg/liter leucine at all glucose concentrations tested, while cells transformed with LEU2-based plasmids did not show this effect of the SPT15-300 allele (Fig. (Fig.3A).3A). In SC−Ura media containing 30 mg/liter of leucine and 5% or 6% ethanol, growth was severely slowed and thus the apparent growth advantage of the SPT15-300 mutant was reduced (Fig. (Fig.3B).3B). However, at 4% ethanol the growth advantage of the mutant was still noticeable, in particular at lower glucose concentrations (Fig. (Fig.3B).3B). When the growth experiments were performed with liquid SC medium containing 20 g/liter glucose and 150 mg/liter leucine, the growth advantage of the mutant SPT15-300 allele was absent (1.1-fold growth improvement; standard deviation, 0.0 [data not shown]). In the presence of 4, 5, or 6% ethanol, the fold growth improvement was limited (1.7 to 1.8; standard deviation, 0.1 [data not shown]). As was the case on solid media, cells transformed with the wild-type and mutant SPT15 alleles on a LEU2 plasmid grew equally well in SC−Leu media with different glucose concentrations, without or with 6% ethanol (Fig. (Fig.3C3C).
These growth experiments illustrate that the enhanced growth of cells with the SPT15-300 mutant allele could be distinguished only in media with limiting amounts of leucine and when expressed from plasmids that do not complement the LEU2 mutation in BY4741 (i.e., URA3- or G418 selection-based plasmids). This implies that the beneficial growth advantage of cells expressing the SPT15-300 mutation is the result of enhanced uptake and/or improved utilization of leucine.
The ethanol sensitivity of S. cerevisiae strains with single-gene deletions (commonly leucine auxotrophic strains) has been determined mainly in rich complex media (12, 16-18, 35, 37, 40). Therefore, a possible effect of ethanol on leucine uptake and/or utilization has not been reported in these global screens. However, impairment of amino acid transport and/or utilization in yeast by ethanol has been described (11). Recently, Hirasawa et al. reported that tryptophan uptake might be inhibited by high concentrations of ethanol (16). Overexpression of TAT2, encoding a high-affinity tryptophan and tyrosine permease (30), yielded yeast cells that acquired a higher tolerance toward ethanol (16). Likewise, the known growth defect of S. cerevisiae leu2 strains (e.g., BY4741) on SC media (8) could be alleviated by overexpression of TAT1 or BAP2 (both encoding amino acid permeases that transport leucine [14, 30]) or by reintroducing LEU2 (8). Several studies have demonstrated that the amount of leucine provided in commonly used synthetic media is limiting for growth of leucine-requiring strains, and authors therefore recommend supplementing synthetic media with at least 400 mg leucine per liter (6, 29).
Based on the growth phenotypes, we decided to reinvestigate the microarray data published by Alper et al. (2), now focusing on uptake and metabolism of leucine. Genes involved in the uptake and degradation of leucine showed differential expression due to the SPT15-300 mutations but also in the presence of increased glucose and ethanol (medium B). The TAT1 gene, which codes for a tyrosine and tryptophan amino acid permease, and the BAP3 gene, which codes for a branched-chain amino acid permease, showed increased expression in cells with the mutant SPT15-300 allele in both media (i.e., media A and B) compared to cells harboring the wild-type SPT15 (Fig. (Fig.4A).4A). BAP2 (coding for another branched-chain amino acid permease) showed a similar expression profile but was not significant at a FDR of 0.005. These three genes code for permease proteins that are able to transport leucine across the plasma membrane (14, 30), and increased expression of TAT1 and BAP2 has been shown to alleviate reduced growth of BY4741 on SC media (8). A majority of the genes involved in leucine utilization and degradation show statistically significant differential expression for the SPT15-300 mutations (a genotype effect, i.e., an effect on gene expression when comparing cells expressing the SPT15-300 allele to those expressing the wild-type allele). ARO10, coding for one of the Ehrlich pathway decarboxylases involved in leucine degradation, is significantly upregulated both by the SPT15-300 mutations and by the presence of increased glucose and ethanol (Fig. (Fig.4B)4B) (38). Additionally ADH3, ADH5, and SFA1, coding for alcohol dehydrogenases, show upregulated expression in the SPT15-300 mutant compared to wild-type SPT15 cells, suggesting that higher rates of NADH reoxidation via 3-methylbutanal reduction in the SPT15-300 mutant could account for the increased fitness of cells with the SPT15-300 allele under leucine-limiting conditions. Finally, genes involved in leucine biosynthesis were downregulated by the presence of ethanol (Fig. (Fig.4C).4C). Expression of the ILV5 and LEU1 genes was only slightly changed due to the presence of the SPT15-300 mutations compared to wild-type SPT15 (genotype effect), though the direction of the response was unchanged. As expected, the LEU2 gene showed only background expression, while genes coding for branched-chain amino acid aminotransferases, i.e., BAT1 and BAT2, showed inverse expression correlating with the cells transitioning from logarithmic to growth-arrested phase (10). It is therefore likely that the improved growth of the SPT15-300 mutant under leucine-limiting conditions is due to increased uptake and utilization of leucine (9, 28, 36).
Alper et al. (2) unambiguously demonstrated that gTME is applicable to S. cerevisiae for altering its properties. Unfortunately, the properties of cells with the mutant SPT15-300 allele did not result in increased ethanol-tolerant phenotypes of yeast in rich complex media, but the application of gTME has been reported to improve xylose fermentation in S. cerevisiae (22).
We thank Lisbeth Faldborg for excellent technical assistance, J. M. Thevelein for providing the S. cerevisiae JT20150 strain, and C. P. Kurtzman for kindly donating S. bayanus strain NRRL Y-11845. We are grateful to M. C. Kielland-Brandt and J. Dietvorst for valuable discussions and critical reading of the manuscript.
Published ahead of print on 7 August 2009.