Introduction

Since the first commercial application of genetically modified (GM) plants, the planting and harvesting of GM plants has increased but consumers have not freely accepted GM organisms (GMO; Stewart et al. 2000; Falk et al. 2002; Harlander 2002; Hino 2002; Moseley 2002; Schilter and Constable 2002). In contrast to GM plants, GM microorganisms useful for the production of fermented foods and alcoholic drinks are not commercially available (for reviews, see Dequin 2001; Akada 2002; Pretorius and Bauer 2002). In 1990 and 1994, prior to the commercial application of GM plants, two GM yeast strains were approved by the British Government for commercial application (Aldhous 1990; Hammond 1995; Smith 1998). However, these GM yeast strains have never been used commercially. For both GM plants and microorganisms, the public remains concerned about the possible toxic and allergic effects of the products derived from the foreign genes in the GMOs when taken up into the human body (Lack 2002). In most cases, useful genes are introduced together with plasmid vectors that contain the drug-resistance genes necessary for gene manipulation in Escherichia coli and other microorganisms or plants. The risk posed by the gene products has been evaluated extensively and the drug-resistance marker genes are unlikely to cause problems if digested in the human stomach and gut (Jonas et al. 2001). However, the possibility of gene transfer to other organisms, such as from GM plants to weeds or from GM microorganisms to pathogenic microorganisms, cannot be neglected (Lorenz and Wackernagel 1994; Adam et al. 1999; Gasson 2000; Stewart et al. 2000; Dale et al. 2002; Gaugitsch 2002; Rieger et al. 2002). If drug-resistance genes and bacterial DNA derived from species other than the host organism were not present in the host organism, these potential problems would be eliminated.

In order to remove unwanted DNA sequences from GM yeast, we developed a method to excise unwanted DNA sequences from recombinant industrial yeast strains (Akada et al. 1999). The traditional method for the elimination of unwanted DNA is a counter-selection system using the URA3 marker, a ura3 − yeast strain, and a drug, 5-fluoroorotic acid (5-FOA; Boeke et al. 1984, 1987). Yeast URA3 + strains cannot grow on 5-FOA plates but ura3 − strains can and therefore the loss of the URA3 marker from transformants allows cells grow on 5-FOA plates. By combining homologous recombination with the counter-selection system, introduced plasmid sequences present on chromosomes can be eliminated completely (Rothstein 1991; Sikorski and Boeke 1991). One disadvantage of the URA3/5-FOA system is the requirement of a ura3 − mutation in the host. Industrial yeast strains are usually polyploid and prototrophic. Therefore, counter-selection systems commonly used for laboratory yeast strains cannot be applied to industrial yeast strains. In contrast, our dominant counter-selection system is useful in any kind of yeast strain (Akada et al. 1999, 2002a; Kawahata et al. 1999).

The first example of the use of the dominant counter-selection system in industrial yeast was a FAS2 replacement (Akada et al. 1999). Based on a two-step gene replacement protocol (Rothstein 1991; Sikorski and Boeke 1991), a FAS2 mutant gene was replaced with a wild-type counterpart in a Japanese sake yeast strain in order to improve the flavor of sake. In this case, a self-cloning yeast that contained only yeast DNA but none of the other DNA sequences used during gene manipulations was constructed. Genetically identical mutant yeast strains, which were selected by drug-resistance screening (Ichikawa et al. 1991), have been used commercially for a long time in sake brewing, due to the value of the flavor profile caused by the FAS2 mutation (Akada et al. 2001). The GM FAS2 mutant yeast, which contains a single base change identical to that in the commercially distributed sake yeast, was approved as a self-cloning strain by the Japanese Ministry of Health, Labor, and Welfare in April 2001. This is the first GM microorganism approved in Japan. The yeast strain, which does not need to be designated as GM, is freely available to the public. As shown in this case, self-cloning is practical and safe and is applicable to the commercial food and drink market without the concerns caused by the presence of drug-resistance genes and bacterial DNA sequences.

If self-cloning is the most direct way to the commercial application of GM microorganisms, the development of methods to improve microorganisms with their own genes is necessary. In order to confer valuable properties to sake, we utilized the yeast ATF1 gene encoding alcohol acetyltransferase in this study. The overexpression of ATF1 in sake and wine yeast is known to affect the properties of the beverages by increasing specific flavor compounds, especially isoamyl acetate (Fujii et al. 1994; Fukuda et al. 1998; Lilly et al. 2000; Mason and Dufour 2000). These sake and wine yeast strains were constructed by transformation with plasmids containing ATF1 and extraneous DNA. Due to the present situation for GM and self-cloning microorganisms, we planned to construct a self-cloning industrial yeast that overexpressed the ATF1 gene by inserting a constitutive overexpression promoter from the yeast TDH3 gene encoding glyceraldehyde-3-phosphate dehydrogenase (Schena at al. 1991; Velculescu et al. 1997) upstream of the 5′ end of ATF1. The resulting self-cloning yeast contained no extraneous DNA sequences derived from foreign species. The self-cloning strain produced a higher amount of isoamyl acetate, a valuable flavor component in sake.

Materials and methods

Strains and media

Japanese sake yeast Kyokai No. 7 strain was obtained from the Brewing Society of Japan. All strains in this study were derived from Kyokai No. 7. Plasmid integrants were RAK1807, RAK1808, and RAK1809. Counter-selected strains RAK1857, RAK1859, and RAK1861, containing TDH3p-ATF1 but no extraneous DNA sequences, were derived from RAK1809. E. coli DH5α was used for plasmid constructions. The yeast strains were grown at 28 °C in YPD medium (2% glucose, 2% Polypepton, 1% yeast extract, with 2% agar when necessary). Synthetic dextrose medium (SD) contained 0.17% yeast nitrogen base without amino acids and without ammonium sulfate, 0.5% ammonium sulfate, and 2% glucose. YPGal medium contained 2% galactose instead of glucose in YPD. Cerulenin and cycloheximide were purchased from Sigma. Cerulenin was dissolved in ethanol and cycloheximide was dissolved in water and added to the media after autoclaving.

Plasmid constructions

Yeast chromosomal DNA was isolated by the protoplast method using Zymolyase 100T (Seikagaku Kogyo, Tokyo), followed by phenol/chloroform extraction and ethanol precipitation (Philippsen et al. 1991) and was used as a template for all PCR amplifications in this study. Primers (Table 1) were dissolved in sterile water at a concentration of 0.1 µg/µl. PCR was performed in a 25-µl reaction mixture, using KOD Dash DNA polymerase (Toyobo, Osaka, Japan) according to the manufacturer’s instructions. For amplification of the TDH3 promoter, a 0.7-kb DNA fragment was amplified using TDH3-1 and TDH3-2 primers. The PCR was initiated at 94 °C for 1 min, followed by 30 cycles of 94 °C for 20 s, 55 °C for 2 s, and 74 °C for 30 s. The DNA fragment was digested with SalI and XhoI and inserted into the SalI-XhoI sites of pRS305 (Sikorski and Hieter 1989) to form p305TDH3p. A 1.6-kb ATF1 fragment was amplified with ATF1-NCO and ATF1-APA primers using the same PCR program. The fragment was digested with NcoI, which is located at the 3′ terminus of ATF1; and the resulting 1.3-kb fragment was inserted into the NcoI site of p305TDH3p to form p305TDH3pATF1-5. The ATF1 gene was located downstream of TDH3 promoter. For amplification of the 5′ upstream region of ATF1, a 2.0-kb fragment was amplified with ATF1-UPXBA and ATF1-UCSAL primers by the following PCR program: 95 °C for 1 min, followed by 30 cycles of 95 °C for 30 s, 56 °C for 2 s, and 74 °C for 2 min. The 2.0-kb fragment was digested with XbaI and SalI and inserted into the XbaI-SalI sites of p305TDH3pATF1-5 to form p305ATTDH3pATF1. A 3.8-kb XbaI-XhoI fragment isolated from p305ATTDH3pATF1 was inserted into the XbaI-SalI sites of pGG119 vector (Akada et al. 2002a) to form pGG119TDH3pATF1 (Fig. 1).

Table 1 Primers. Restriction sites are in italics
Full size table
Fig. 1
figure 1

Structure of the pGG119TDH3pATF1 plasmid. The pGG119 vector contained the PGKp-YAP1 drug-resistance transformation marker and GALp-GIN11M86 counter-selection marker. TDH3p (glyceraldehyde-3-phosphate dehydrogenase 3 gene promoter) was fused to ATF1ΔC (alcohol acetyltransferase 1 lacking its C-terminus) and the 5′ upstream non-coding sequence of ATF1 was placed upstream of the TDH3 promoter. The single SphI site within ATF1 was used for restriction enzyme digestion prior to integrative transformation

Full size image

Yeast manipulation

Yeast transformation was performed according to the simple lithium acetate/polyethylene glycol method (Chen et al. 1992). pGG119TDH3pATF1 was digested with SphI and used to transform Kyokai No. 7. The transformation mixture was spread on SD plates containing 1.0 µg cerulenin/ml. The cerulenin-resistant colonies were streaked on SD plates containing 0.5 µg cycloheximide/ml as a double-check, because of the dual resistance caused by the PGKp-YAP1 marker (Akada et al. 2002b). For counter-selection, transformants were grown at 28 °C for 1 day in liquid YPGal medium and then cells were spread on YPGal plates at approximately 103–104 cells/plate. Colonies that grew on YPGal plates after 36–48 h incubation at 28 °C were picked.

Southern blot analysis

Chromosomal DNA was digested with AflII, run on an agarose gel, and transferred to a Nylon membrane. A 1.2-kb NcoI fragment containing ATF1 from p305TDH3pATF1-5 and BamHI-digested pBluescript KS+ were used as probes. Probe preparation and hybridization were performed as described by Akada et al. (1999), using the DIG system (Roche Diagnostics, Germany) with digoxigenin.

Sake brewing test

Sake brewing was performed at 10 °C using a total of 500 g rice and 675 ml water. Fermentation was monitored by the loss of weight caused by CO2 production. The general components of sake were analyzed by a standard method established by the National Tax Administration Agency (Brewing Society of Japan 1993). The concentrations of ethyl caproate and isoamyl acetate were measured as described by Akada et al. (1999).

Results

Construction of TDH3p-ATF1 plasmid and self-cloning strategy

For the self-cloning of TDH3-ATF1, we constructed pGG119TDH3pATF1 (Fig. 1). The pGG119 vector contained a drug-resistance marker, PGKp-YAP1, for transformant selection, and a counter-selection marker, GALp-GIN11M86, for the selection of plasmid loss (Akada et al. 2002a). The vector was inserted with a 3.8-kb fragment containing the 5′ upstream region of ATF1 (5′-up) and TDH3p-ATF1ΔC, which contains a deletion of the C-terminal sequence of ATF1 for confirmation of correct integration by PCR.

The self-cloning strategy of TDH3p-ATF1 is shown in Fig. 2A. The pGG119TDH3pATF1 plasmid was linearized by cutting within the ATF1 open reading frame and was integrated into the ATF1 locus of Kyokai No. 7. From the integrant, strains that lost the plasmid sequences were identified by counter-selection using GALp-GIN11M86 as indicated by growth on galactose plates. Loss of the plasmid occurred through homologous recombination either between the 5′ upstream sequences of ATF1 or between the TDH3p downstream sequences of ATF1. Recombination between the 5′ upstream regions resulted in a TDH3p-ATF1 strain, but that between the downstream ATF1 resulted in reversion to a wild-type strain.

Fig. 2A, B
figure 2

Self-cloning strategy for promoter insertion by a two-step protocol. A pGG119TDH3pATF1 was digested with SphI and inserted into the ATF1 locus of a sake yeast strain. Plasmid integration resulted in the insertion of plasmid DNA sequences between the tandem repeats of ATF1. Homologous recombination between the TDH3p-downstream ATF1 sequences resulted in a wild-type ATF1 gene, but that between the 5′ upstream sequences resulted in TDH3p-ATF1 without any extraneous DNA sequences. Arrows indicate the sites of primers used for PCR analysis (Fig. 3). ATF1ΔC does not contain the primer site and thus only the correct integrants can produce a 2.3-kb TDH3p-ATF1 fragment. B Wild type, integrant, and counter-selected strains were streaked on cerulenin and cycloheximide plates. The integrant, but not the counter-selected strains (sectors 1–6) grew on the drug-resistance plates, indicating the loss of the transformation marker by counter-selection

Full size image

Construction of a self-cloning TDH3p-ATF1 yeast

pGG119TDH3pATF1 was digested with SphI (Fig. 1) and used to transform the Kyokai No. 7 strain. Several colonies that grew on cerulenin plates (approx. 100 colonies/µg) were picked onto cycloheximide plates for the double selection (Akada et al. 2002b; Fig. 2B). Three strains (RAK1807, RAK1808, RAK1809) were selected as the correct plasmid integrants after Southern blot analysis using the ATF1 probe, which showed 23-kb bands in addition to 12-kb bands, because the Kyokai No. 7 strain is diploid (Figs. 2, 3A)

Fig. 3A, B
figure 3

Southern blot analysis and PCR detection of TDH3p-ATF1. A Chromosomal DNAs were isolated from Kyokai No. 7 strain, plasmid integrants, and counter-selected strains, digested with AflII, and subjected to agarose gel electrophoresis. Integrants showed a 23-kb band in addition to a 12-kb band when hybridized with the ATF1 probe, but counter-selected strains (1857, 1858, 1861) derived from the RAK1809 integrant showed only the 12-kb band. The bacterial vector DNA probe (pBluescript KS +) hybridized to the 23-kb band but not to the chromosomal DNAs from Kyokai No. 7 or the counter-selected strains. B PCR amplification with primers TDH3-1 and ATF1-APA using template DNAs from the RAK1809 integrant and counter-selected strains produced a 2.3-kb band in strains RAK1857, RAK1858, and RAK1861 but not in RAK1859, indicating that the three counter-selected strains contained TDH3p-ATF1

Full size image

From the RAK1809 integrant, counter-selected strains were selected by growth on YPGal plates. Colonies that grew on the galactose plates were streaked onto cerulenin and cycloheximide plates (Fig. 2B). All selected strains lost resistance to cerulenin and cycloheximide, indicating that these strains lost the PGKp-YAP1 marker of the plasmid. These strains were either TDH3p-ATF1 or ATF1, as shown in Fig. 2A. To select TDH3p-ATF1 strains, PCR was performed with TDH3-1 and ATF1-APA primers and chromosomal DNA isolated from counter-selected strains. This PCR could not produce a product from pGG119TDH3pATF1 due to the absence of the annealing site of ATF1-APA. However, a 2.3-kb fragment was produced from integrants because correct integration generated TDH3p-ATF1, which contained the annealing sequence of the ATF1-APA primer. Among four counter-selected strains, three strains, RAK1857, RAK1858, and RAK1861, produced a 2.3-kb band, indicating that these were TDH3p-ATF1 strains (Fig. 3B). The ATF1 probe showed only the 12-kb band in hybridizations with these three strains. The 23-kb band was not present in the three integrants. The 0.7-kb difference between the inserted TDH3 promoter and wild-type ATF1 was not obvious in this Southern blot analysis. Loss of the extraneous plasmid DNA sequences in these strains was confirmed by Southern blot analysis using a pBluescript probe to detect the E. coli vector backbone of pGG119 (Fig. 3A). The pBluescript KS+ probe hybridized to the 23-kb band of the integrants but showed no hybridization to the chromosomes of counter-selected strains or the Kyokai No. 7 strain, indicating that there were no bacterial vector sequences in these strains. From these results, we concluded that the self-cloning TDH3p-ATF1 strain construction was successful.

Sake profiles

Laboratory-scale sake fermentation was performed with wild-type Kyokai No. 7, integrant RAK1809, and self-cloning RAK1857 strains. The fermentation profiles are shown in Table 2. The three strains showed similar fermentation profiles, except for acetic acid and isoamyl acetate. The integrant and self-cloning strains produced three and five times more isoamyl acetate than the wild-type strain, respectively, and less acetic acid, indicating a higher alcohol acetyltransferase activity.

Table 2 Sake profiles
Full size table

Discussion

We constructed a sake yeast strain that overexpresses the ATF1 gene by placing a yeast constitutive overexpression promoter, TDH3p, upstream of the 5′ end of ATF1. The vector sequences used for gene manipulations were removed completely. The resulting strain contained only a 0.7-kb TDH3 promoter sequence derived from the same Kyokai No. 7 strain upstream of ATF1. The strain construction process can be called “self-cloning”, which means cloning of its own DNA (Walgate 1979). The self-cloning TDH3p-ATF1 strain produced a higher amount of isoamyl acetate due to the overexpression of the ATF1 gene, as shown previously in winery, distillery, and sake yeast strains, in which ATF1 was overexpressed by transforming shuttle plasmids (Fujii et al. 1994; Lilly et al. 2000). The originality of this study is the self-cloning process for constructing commercially applicable GM yeast strains that overexpress useful genes.

The concept of self-cloning is important for the commercial use of GM microorganisms. Since the revision of the guidelines for safety assessments of GM organisms issued by the Japanese Ministry of Health, Labor, and Welfare, which became effective from April 2001 (Hino 2002), this self-cloning type of GM organism was eliminated from food safety assessments. Any GM microorganisms that are approved as self-cloning are not treated as GMO in Japan (Akada 2002). Actually, self-cloning is reasonable for the production of GM foods, because there is no risk of problematic gene transfer from GM foods to other microorganisms living in the human or animal body. Therefore, self-cloning gene manipulation is now practical for the commercial application of GM organisms.

A two-step gene replacement protocol was usually used for gene replacement between mutant and wild-type alleles in laboratory yeast strains by using a counter-selection system such as the URA3/ura3/5-FOA and LYS2/lys2/aminoadipate systems (Boeke et al. 1984; Sikorski and Boeke 1991). For industrial yeast strains, such gene replacement methods requiring recessive mutant yeast are not applicable. Using the dominant counter-selection system, FAS2 point-mutant genes were successfully replaced with the wild-type FAS2 in sake yeast to provide a valuable flavor profile in sake (Akada et al. 1999; Aritomi et al. 2004). In addition to the introduction of these point-mutation alleles, to the best of our knowledge, this report provides the first example of promoter insertion using the two-step protocol. The self-cloning procedure and the resulting TDH3p-ATF1 strain, which confers an improved flavor profile in sake, will be applicable to the Japanese sake market and also to the world GM market.