Sake, a Japanese alcoholic beverage, is considered as an item of luxury in Japan. It is in demand for various products based on diverse consumer needs. Consumers are particularly sensitive to the taste, aroma, and alcoholic content, which are influenced by the yeast strain that plays a major role in fermentation. Sake is produced from rice and water by fermentation using the koji mold, Aspergillus oryzae (Ahlb.) Cohn, and sake yeast, Saccharomyces cerevisiae (Desm.) Meyen.
Sake yeasts are extensively studied owing to their impact on product quality and process efficiency. Several researchers have attempted to breed sake yeasts with favorable characteristics, such as heightened resistance to ethyl alcohol and enhanced flavor production, to improve the overall product quality (Kitagaki & Kitamoto, 2013). Additionally, several wild strains of S. cerevisiae are found in nature (Sampaio & Gonçalves, 2008). Yeast strains used in brewing are ubiquitous, and some natural yeast strains may impart unique flavors and characteristics to sake. Therefore, sake produced using S. cerevisiae isolated from regional sources may hold special appeal for consumers. Furthermore, natural isolates of S. cerevisiae can have more desirable characteristics for sake brewing than sake yeast strains produced industrially.
In this study, we isolated a natural strain of S. cerevisiae from litter sample collected at Muramatsu Park in Gosen City, Japan. Cerasus × subhirtella ‘Hozaki-higan-yaezakura’ is a rare tree that grows in this park. Furthermore, this park is renowned as a "flower park" with beautiful seasonal blooms. Therefore, we isolated yeast strains that can be candidate for breeding sake yeast in this park, with the dual aim of attracting interest from consumers and harnessing novel microbial resources.
As shown in Fig. 1A, between 2020 and 2022, various flowers, plants, and plant litter were collected using sterilized instruments. According to our novel scheme, we screened for natural yeasts using the loop-mediated isothermal amplification (LAMP) assay based on the target gene PHO3 (Kuribayashi et al., 2014), which exists in natural yeast but not in the current industrial sake yeasts. The LAMP assay is a more rapid and simple method compared to the classic plating method and/or PCR method, facilitating the facile isolation of natural yeast. Using the LAMP method, a yeast-like microorganism harboring the PHO3 gene, namely HG-3, was isolated from the plant litter of ‘Hozaki-higan-yaezakura’. To identify and characterize the strain, the nucleotide sequence of the D1/D2 domain of the 26S rDNA of HG-3 was analyzed, as previously described by Kawahata et al. (2007). Universal primer sets (NL-1: 5′-GCATATCAATAAGCGGAGGAAAAG-3′) and (NL-4: 5′-GGTCCGTGTTTCAAGACGG-3′) were used to amplify the 26S rDNA region from the isolated DNA. The 26S rDNA sequence of HG-3 was identical to that of typical strains of S. cerevisiae species (DDBJ accession no. LC797988)
Fig. 1. Screening and breeding scheme of Saccharomyces cerevisiae isolated from the Muramatus Park in Gosen city. A: Screening using loop-mediated isothermal amplification (LAMP) assay and identification of a yeast strain. B: A breeding program for a hybrid strain for sake brewing.
Almost all industrial sake yeasts, including the Kyokai yeast strains, a superior sake yeast variety used in industrial sake brewing, can be distinguished from wild sake yeasts using the detection of acid phosphatase activity in a high phosphate medium (Mizoguchi & Fujita, 1982). The Kyokai yeast strain colonies do not show acid phosphatase activity owing to a mutation in the PHO3 gene (Kuribayashi et al., 2014). They do not possess sporulation ability (Suizu et al., 1996). In contrast, wild sake yeast strains demonstrate acid phosphatase activity and form spores. As these characteristics of HG-3 are consistent with those of wild sake yeast, we concluded that HG-3 does not belong to the modern Kyokai strain group (Table 1). These findings suggest that HG-3 is a natural isolate, distinct from industrial sake yeast strains.
Table 1. Physiological characteristics of the industrial yeast strain Kyokai no. 901 (K901), isolated yeast strain HG-3, haploid strain KH-1, and HG-3-F2.
| Physiological characteristics | Yeast strain |
| K901 | HG-3 | KH-1 | HG-3-F2 |
| Acid phosphatase activitya | − | + | − | + |
| Killer factorb | − | − | − | − |
| Sporulationc | − | + | − | − |
+, positive; −, negative
a The detection of the acid phosphatase activity was carried out in accordance with Mizoguchi and Fujita (1982).
b The secretion of killer toxin was examined using the halo test. The assay was preformed using K901 cells as indicator cells.
c For sporulation tests, yeast strains were cultured on YPD plates at 28 °C for 3 d. Then, they were transferred to solid sporulation media (0.5% sodium acetate and 2% agar) and cultured at room temperature for 1-2 days. The detection of sporulation on each plate was observed using a light microscope.
Next, to introduce the ability of sake brewing, this natural yeast was cultured using the mass spore-cell/cell-cell mating common method (Lindegren & Lindegren, 1943) with a sake yeast haploid, KH-1 (Table 1), which was isolated from Kinshihai shuzo sake yeast strain. For efficient mating, an arginase-deficient mutant was isolated from the HG-3 strain using the positive selection medium containing canavanine, arginine, and ornithine (CAO medium) (Kitamoto et al., 1993). Then, by following the breeding program, a hybrid yeast strain, HG-3-F2, was constructed (Fig. 1B; Table 1).
To examine the fermentation ability, a sake brewing test was performed with the wild strain HG-3, the hybrid strains, and Kyokai no. 901 (K901) with 200 g of total rice. Rice, koji, and water were added to the sake mash, according to the method described by Namba et al. (1978). A yeast extract peptone dextrose (YPD) medium was used for yeast preculture. The fermentation temperature was maintained at 15 °C, and the amount of CO2 evolved was measured daily. After fermentation was performed for approximately 10 d, the sake mash was centrifuged, and the supernatant was separated and processed to obtain sake. The CO2 evolution rate of HG-3 was 52% lower than that of K901 (Table 2). The amounts of ethanol, flavor components, higher alcohols, and volatile esters produced by HG-3 were significantly lower than those produced by K901. These results suggest that HG-3 may possess desirable characteristics for the production of sake with reduced alcohol content.
Table 2. Analysis of HG-3-F2 for sake brewing.
| Yeast strain | CO2 evolution (g) | General properties | Flavor components (mg/L) |
| Sake meter | Alcohol (% (v/v)) | Acidity (mL) | Amino acidity (mL) | Isobutyl alcohol | Isoamyl alcohol | Isoamyl acetate | Ethyl caproate |
| Evaluation of HG-3-F2 for small-scale brewing (200 g of total rice) |
| K901 | 60.0 | −22.4 | 17.6 | 3.9 | 1.9 | 73.9 | 186.8 | 3.1 | 2.3 |
| HG-3 | 28.8* | −85.7* | 9.7* | 4.2 | 2.0 | 51.8* | 139.5* | 1.1* | 0.9* |
| HG-3-F2 | 59.2 | −21.4 | 17.4 | 3.3† | 2.0 | 46.9† | 169.5† | 3.5 | 1.5 |
| Preliminary test for industrial-scale brewing (200 g of total rice) |
| KH-1 | 59.3 | −40.9 | 17.0 | 2.9 | 1.6 | 76.2 | 130.1 | 3.8 | 1.0 |
| HG-3-F2 | 58.6 | −42.8 | 16.6‡ | 3.0 | 1.5 | 54.6‡ | 140.9 | 3.9 | 1.4‡ |
| Industrial-scale brewing (300 kg of total rice) |
| HG-3-F2 | ND | −1.8 | 16.6 | 2.2 | 1.5 | 33.6 | 142.7 | 2.6 | 2.1 |
For the tests using 200 g of total rice, the values are the average of three independent sake-brewing experiments. The product lots of raw materials (rice and koji) differed between the two experimental systems.
Industrial-scale (300 kg) production was conducted once only; hence, these results were not evaluated statistically.
ND: no data.
* Significant difference, HG-3 compared with K901 (Student's t-test, p < 0.05).
† Significant difference, HG-3-F2 compared with K901 (Student's t-test, p < 0.05).
‡ Significant difference, HG-3-F2 compared with KH-1 (Student's t-test, p < 0.05).
In contrast, the HG-3-F2 strain showed higher alcoholic fermentation than K901, and its desirable ability of sake brewing produced lower acidity and flavors (isobutyl alcohol and isoamyl alcohol, such as higher alcohols) compared with K901 (Table 2). Taste, flavor, and overall quality were determined using four-panel lists, following which the hybrid strain, HG-3-F2, was confirmed to be suitable for sake brewing (data not shown).
To evaluate whether the hybrid strain HG-3-F2 was indeed suitable for commercial application, we performed sake brewing tests with this strain in a sake brewery (Kinshihai Shuzo Co., Ltd.). First, we confirmed that the HG-3-F2 strain exhibited lower alcohol content and isobutyl alcohol than the KH-1 strain, with higher levels of ethyl caproate than KH-1, producing a smoother flavor. The HG-3-F2 strain and this brewery's haploid strain KH-1 thus exhibited different brewing properties for 200 g of total rice (Table 2). Next, we performed an industrial-scale brewing test by fermenting 300 kg of total rice with HG-3-F2 (Junmai-shu grade; Gohyakumangoku and Koshiibuki as material rice, rice polishing rate of 65%). In this scaled-up test, the alcohol content and flavor profile necessary for commercialization were successfully obtained (Table 2). The hybrid strain HG-3-F2 is therefore suitable for industrial-scale sake brewing.
In conclusion, this study is the first in Japan to successfully isolate a candidate yeast strain for sake production from nature using the LAMP method. Further, we successfully constructed sake yeast via yeast mating. Our serial procedure may be useful for sake brewers because it allows for a simple yeast isolation and enables the conventional production of sake even with isolated yeasts that have a low alcohol-producing ability. Although more detailed analyses on HG-3 and HG-3-F2 are required, we expect these strains to contribute to high-quality sake production in Gosen City, Niigata Prefecture, Japan.
