Research advances on sake rice, koji, and sake yeast: A review

Abstract Sake is the national alcoholic beverage of Japan, and its history can be traced back more than 1300 years. With the development and maturity of the sake‐brewing technique, a unique flavor and taste gradually formed, which led to its wide use in Japan and internationally. This paper reviews and discusses the research advances of sake rice, koji, and sake yeast. The various enzymes and involved genes of microbes in the rice koji, and the separation/breeding of sake yeasts are expounded particularly. Moreover, the fields where further research is required are presented. Therefore, this review presents recent comprehensive research details of sake's ingredients and the involved study perspectives.

but also by environmental factors encountered during cultivation.
Yamadanishiki cultivated in Japan's Hyogo Prefecture in 1923, was the most popular rice for sake making, and was widely planted. Its best producer was in Hyogo Prefecture, Japan. Yamadanishiki has a higher white core expression rate (WCE) and is lower in protein and fat compared with cooking rice cultivars. Okada et al. (2017) reported that the 100-grain weight and WCE of Yamadanishiki were 1.30 and 7.74 times higher than those of Koshihikari (cooking rice cultivars), respectively. While the Yamadanishiki alleles at 16 QTLs contributed to larger grain size, two major QTLs that are essential for regulating both 100-grain weight and grain width were harbored in the same regions on chromosomes 5 and 10. The QTLs associated with grain size also play an important role in the formation of white core, and the formation of white core does not depend on the grain-filling speed (Okada et al., 2017). The Gohyakumangoku, which had large consumption among Japan's sakes, was surpassed by Yamadanishiki in 2001. Due to its large white core and frangibility, when it was ground by 50%, this rice was not suitable for daiginjo brewing. Gohyakumangoku rice had always been famous for its savory and elegant wine quality. Due to its strong cold tolerance, it was cultivated all over Japan. The Miyamanishiki rice was named since its white core looked like the white snow of the famous Alps in the Nagano Prefecture, and its production ranked third in Japan. Sake made from Miyamanishiki rice is fresh and has a pleasant fragrance (Yamamoto, 2017).
After harvest, the brown rice with large grain, white core, and without cracks is selected for grinding, and the resulting ground rice is called polished rice. The milled rice ratio refers to the proportion of polished rice in the original brown rice. Junmai sakes are generally divided into three grades according to the different milled rice ratio: The first grade is junmai daiginjo, which is brewed by pure rice with a milled rice ratio of 50% or less; the second is the junmai ginjo-syu with a milled rice ratio of 60% or less; and the third grade is the junmai-syu without the restriction of milled rice ratio. Since the protein content is higher in the outer layer than in the center of rice grains, and protein constitutes approximately 68% of the dry weight of brown rice grains, and 4%-6% of the dry weight of sake-making rice grains with a 70% polishing ratio. If oligopeptides, amino acid, and unsaturated fatty acids are too high, negative effects, such as excessive coloration and/or unpleasant taste, may result. Hashizume et al. reported that five bitter-tasting peptides, consisting of six to thirteen amino acid residues, were detected in sakes by RP-HPLC and gel permeation chromatography (Hashizume, Okuda, Numata, & Iwashita, 2007). Peptide no. 17, whose C-terminus was histidine, had the lowest threshold value (0.06 μM) and was described as having a particular "zatsumi" taste, and the C-terminal residue of peptides for peptides nos. 13, 18, and 20 was determined to be proline, which has a hydrophobic side chain and conveys a negative taste with a bitter note since it is generally assumed that the bitterness of peptides is closely related to their hydrophobicity (Ney, 1971).
Moreover, Hashizume et al. (2007) also reported that the ginjyo-type sake made from highly polished (40%) rice grain had a lower level of bitter-tasting peptides than the jyunmai-type sake made from 70% polished rice grains (Hashizume et al., 2007).  (Yoshizawa, 1999). Therefore, the quality of the resulting sake is suggested based on the local statement in the Hyogo Prefecture as: "the whiter the rice, the better the sakes." Furthermore, Okuda et al. described that steamed sake rice grains, the starches which were low in amylose and had abundant short-chain amylopectin showed high viscosity, little retrogradation, and low gelatinization temperatures. Furthermore, the enzyme digestibility of steamed rice grains in sake mash can be predicted by the mean air temperature during grain-filling (one month after plant heading), and the prediction formula: y = −1.42x + 49.30 (R 2 = 0.748, n = 69), where y represents the enzyme digestibility of steamed rice grains, and x represents the mean air temperature during grain-filling. These results can help to control sake production at an early stage of the sake-making season (Okuda, Aramaki, Koseki, Inouchi, & Hashizume, 2006;Okuda, Aramaki, Koseki, Satoh, & Hashizume, 2005).

| RI CE KOJ I
"The first step is making koji; the second is making yeast starter, and the third is brewing (mash)," remained the most refined summary of the sake-brewing process of the Japanese for thousands of years, indicating that koji played a central role in the brewing of sake. 20% of rice from the raw material is soaked in water, steamed, inoculated with Aspergillus, and cultivated to obtain koji. Koji had three functions: One is to provide an enzyme source for the mash to dissolve and decompose the starch, protein, and fat in the rice; the second is to produce vitamins and amino acids, to offer nutrition for the sake yeast; and the third plays a important role in the flavor of sake (Akiyama & Zhou, 2002).
Aspergillus oryzae is the source of amylase and was adopted in Japanese koji making, which is a crucial fungus applied in traditional Japanese fermented food and beverage production. In 2005, its genome was sequenced (Machida et al., 2005  Ito & Nessa exposed A. oryzae spores to γ-ray mutagenesis and reported that the amylase production capacity of the mutant A. oryzae increased by 2-to 5-fold, corresponding to the irradiation doses. In addition, the spore size and DNA content of the highyield mutant strain increased accordingly (Ito & Nessa, 1996). Soon afterward, Spohr et al. discovered that the α-amylase production ability of A. oryzae to related to its morphological structure. Strains with dense mycelia presented a higher α-amylase production capability (Spohr, Carlsen, Nielsen, & Villadsen, 1997). And then, Agger, Spohr, & Nielsen (2001) reported that the biomass concen- oryzae. They reported that the contribution of amyA was lower than that of amyB and amyC. The reason for the large amount of amylase production by A. oryzae was the duplication of amyB and amyC genes.
The cell walls in plant cells are able to sequester the starch to prevent it from being hydrolyzed, while the enzymes that hydrolyze cellulose and hemicellulose in the cell wall are less expressed in A.
oryzae. To improve the material utilization rate and reduce the slag in the brewing process, Sato et al. purified the hemicellulose-degrading xylanase isoenzyme XynG2 from A. oryzae RIB218, which retains most activity after exposure to 80% alcohol for 30 min. When it is applied as an exogenous enzyme in the wine-making process, the output of sake increased by 4.4% and the vinasse decreased by 4.6% (Sato, Fukuda, Zhou, & Mikami, 2010). In addition, the cellulase Cel-2 isolated from solid medium by Yamane et al. can promote the decomposition of steamed rice and significantly improve the utilization of raw materials during the brewing process (Yamane, Fujita, Izuwa, et al., 2002;.
During the process of koji making, the reproduction of Aspergillus is affected by temperature and moisture. Hence, it is particularly significant to maintain constant temperature and moisture. Since the temperature within the rice is higher than its outside temperature, the reproduction of Aspergillus may be uneven. Prior to the mechanization era, the control of temperature and moisture mainly relied on manpower to mix rice koji, which required hard work. The wooden tool used for koji sometimes contained a substance called TCP (2,4,6-trichlorophenol) originated from wooden tools treated with fungicides, which was a precursor of 2,4,6-trichloroanisole (TCA), and the reason for the musty/muddy smell of sake (Miki, Isogai, Utsunomiya, & Iwata, 2005). Along with the advancement of technology, automatic koji machines were introduced instead of manpower to control the temperature of the product by supplying air with specific temperature and humidity. With the gradual increase in the curved box volume, the koji-making efficiency was improved and the labor needed was decreased .

| SAKE YE A S T
Sake yeast refers to S. cerevisiae strains with different characteristics than other strains and with good suitability for sake brewing. Sake yeast can produce a higher concentration of ethanol than laboratory yeast during the brewing process, which is likely because the buoyant density and stress tolerance of the septic yeast cells in the stationary phase are lower than those in the laboratory yeast cells.
After the cells stopped growing, it was difficult for sake yeast cells to enter a static state (Urbanczyk et al., 2011). Watanabe, Wu, et al. Japanese cultivates a variety of yeasts with different characteristics (Table 1) (Gu, 2016;Kitamoto, Oda-Miyazaki, Gomi, & Kumagai, 1993;Kuribayashi, Sato, Joh, & Kaneoke, 2017). For instance, the K-6 (Kyokai No. 6 yeast, similarly hereinafter) and K-7 are officially distributed by the Brewing Society of Japan and produce their specific fragrance at low temperatures (10-12°C). K-9 is suitable for the production of high-grade sake, while K-10 is suitable for brewing low-acidity wine. K-11 yeast has strong resistance to ethanol. K-7 and K-9 are the most widely applied strains in sake production. The K-7 was isolated in 1946, showed good fermentation performance, and was commonly employed to make ordinary sake.
K-9 was isolated in 1953 and was often used to brew ginjo-syu due to its fruity and refreshing taste. K-601, K-701, and K-901 belong to nonfoaming yeasts where high foam does not form during the fermentation process. Hence, the space of the tank can be fully utilized, which improves productivity.
The Japanese sake industry attaches great importance to the breeding of sake yeast. Sake yeast breeding research was firstly Eht1 and Eeb1 to form ethyl caproate. Today, this strategy to breed an ethyl caproate-overproducing sake yeast is widely utilized to breed sake yeasts for the brewing of ginjo sake (Kitagaki & Kitamoto, 2013).
It is well known that sake fermentation with a nonfoaming yeast will decrease the production cost, because the volume of the required sake tank can be saved. Akiyama et al. isolated nonfoam-forming yeast strains (S-127, S-139) from sake mashes and reported that nonfoam-forming yeasts do not stick to foams though they had some weakness in attenuating power, high productivity of acid, and in the quality of production (Akiyama, Sugano, Kumagai, Saito, & Shimizu, 1969). Momose, Iwano, and Tonoike (1969) reported that the electrostatic binding of the cell walls of yeasts and lactic acid bacteria is responsible for the coflocculation of both yeasts and lactic acid bacteria. Shimoi et al. (2002) confirmed that the factor responsible for the foam formation of the mash by sake yeast is the cell wall protein Awa1, which is rich in serine and threonine residues, contains many repetitive sequences, and has the cell surface hydrophobicity to generate bubbles.  (Arikawa et al., 2000) Improved isoamyl acetate productivity K7-VPA LS K−7/ Growth in a valproic acid (VPA) containing medium 2017 (Tomimoto et al., 2017) Improved isoamyl acetate productivity and lower isoamyl alcohol content Ethyl carbamate (ECA) naturally occurs in fermented foods and beverages. It is spontaneously produced by the reaction between urea and ethanol, and is suspected to be a carcinogen at high doses in animal tests. Therefore, ECA levels in food products are to be reduced. Kitamoto et al. first succeeded in breeding a nonurea-producing diploid sake yeast mutant by deleting the arginase gene CAR1 on two chromosomes of sake yeast (K-9), via genetic engineering.
Then, they brewed sake without no urea with the gene CAR1 deletion homozygous mutant, and no ECA was detected in the resulting sake, even after storage for 5 months at 30°C. The results indicated that ECA in sake originates mainly from urea, which is produced by arginase. Then, Kitamoto et al. (1993) used the CAR1 mutant to develop a new medium for the positive selection of CAR1 mutants, and many arginase-deficient mutants could be easily isolated not only from a laboratory haploid strain, but also from sake yeasts and wine yeasts. No urea was detected in sake brewed with these isolated mutants, and no ECA formed during storage at 70°C for 10 hr. Most mutants had virtually the same fermentation profiles as their parent strain. The medium for the positive selection was as follows: Orn medium [0.17% yeast nitrogen base, 5 mM ornithine, 2% glucose] and Arg medium [0.17% yeast nitrogen base, 5 mM arginine, 2% glucose] were used for the confirmation of CAR1 mutants. CAO medium [0.17% yeast nitrogen base, 10 mg/L canavanine, 5 mM ornithine, 1 mM arginine, 2% glucose] was used for the positive selection of CAR1 (Kitamoto, Oda, Gomi, & Takahashi, 1991;Kitamoto et al., 1993). The breeding of nonurea-producing sake yeast was continuous until now. The gene CAR1 encodes arginase and is located upstream of the FAS2 locus on chromosome XVI of S. cerevisiae. Kuribayashi et al. (2017) isolated nonurea-producing sake yeasts from a FAS2-G3748A (G1250S) mutant using a CAO-selective medium after selection of FAS2 mutants with cerulenin. The resulting double mutants with strong ethyl caproate-producing phenotype (As high as ~ 3.5 times that of the control) does not produce urea during sake fermentation, indicating it as a suitable candidate for the production of safe and high-quality sake.
Isoamyl acetate has a banana-like flavor and is one of the main ester flavors that determine the quality of ginjo sake. Therefore, the main focus was to improve the production of isoamyl acetate. Two major routes have been used for the biosynthesis of isoamyl alcohol. One is via α-keto isocaproate in the L-leucine synthesis pathway from glucose. The other is the so-called Ehrlich mechanism from L-leucine in rice and koji through transamination. In sake fermentation, the biosynthesis of isoamyl alcohol could either be inhibited or decreased by the accumulated L-leucine. Ashida et al. reported a procedure to obtain the mutants of S. cerevisiae that produce sufficient isoamyl acetate, isolated from sake yeast RIB6002 and RIB6006. They used 5,5,5-Trifluoro-DL-leucine as L-leucine analog for the isolation to eliminate the feedback inhibition by accumulated L-leucine. The concentration of isoamyl alcohol that was achieved with these mutants increased by about three or four times compared with the wild strain and reached about 666-881ppm. A sufficient concentration of isoamyl acetate (25-29 ppm) was consequently obtained in the sake fermentation test of the mutants and wild strain (Ashida, Ichkawa, Suginami, & Imayasu, 1987). Arikawa, Yamada, Shimosaka, Okazaki, and Fukuzawa (2000) constructed diploid mutants from sake yeast by EMS treatment and selected three strains that produce a higher level of isoamyl acetate (15.5-17.8 mg/L), which were over twofold higher than that by the parent K901. In recent years, the screening of drug-resistant mutants of sake yeast strains has been an effective method for the creation of superior strains. Tomimoto et al. reported that they succeeded to isolate a valproic acid (VPA)-resistant mutant (K7-VPA LS ) of K-7 using a VPAcontaining medium. The mutant was significantly more resistant to not only VPA-induced cell death but also ethanol compared with the parent strain. Furthermore, the major characteristics of sake brewed by strain K7-VPA LS were lower amino acidity and higher isoamyl acetate content without increased isoamyl alcohol in comparison with/ K7. Moreover, taste sensor analysis showed that the resulting sake has milder sourness and higher saltiness or richness than K-7-brewed sake. High isoamyl acetate production may be connected with a deficiency in phosphatidylinositol, which directly inhibits alcohol acetyltransferase, an enzyme responsible for isoamyl acetate synthesis (Tomimoto, Akao, & Fukuda, 2017).
The selection of S. cerevisiae by haploid hybridization was also commonly employed by researchers. Sake-brewing yeast is a key microorganism in the production process of sake. The accumulation of alcohol in the fermentation process exerts toxic effects on yeast cells, thus inhibiting cell growth and metabolism as well as preventing the production of higher concentrations of alcohol.
Kurose et al. treated haploid yeast/K701 with ethanol to screen for a haploid strain with resistance to 18% (vol) ethanol. A diploid strain was formed via hybridization, which could produce a large amount of fruity isoamyl acetate. This hybrid yeast was used in a sake-brewing experiment at laboratory scale. A higher survival rate was achieved at the later stage of sake brewing, and the yields of isoamyl acetate and ethyl hexanoate were twice more than that of the parent strain (Kurose, Asano, Tarumi, & Kawakita, 2000).
Protoplast fusion technology is an important strain improvement technology that originated in the 1960s. It refers to the process of parental cell fusion after removing the nucleus to obtain fused cells via exchange and recombination among genomes. The sake produced by S. cerevisiae YM39 contained lower amino acid and lower succinic acid contents than K-701, while the malic acid and isoamyl alcohol contents were higher than in K-701. Iwase et al. employed the protoplast fusion technology to fuse YM39 and K-701 and isolated two stable tetraploid fusions. Wine-making experiments showed that the sake brewed by fusion had lower amino acid acidity than K-701, whose isoamyl acetate content was higher than that of the two parents (Iwase et al., 1995). Through protoplast fusion, yeast can also achieve better traits, such as high temperature resistance, improved maltose assimilation ability and pleasing taste of sake. Thus, it had good prospects in the wine industry.
"Hineka" refers to a bad smell of aged sake produced after stor- were lower than the threshold value, while the concentrations of DMTS in the stored sake were greatly decreased. Furthermore, the sake components were almost identical as that of ordinary sake. Thus, the odor development during the sake storage was reduced and the quality of the resulting sake could be improved without affecting the properties of sake (Ikeda et al., 2018).

| CON CLUS I ON AND PER S PEC TIVE S
Sake is the national alcohol beverage of Japan and is welcomed by consumers due to its refreshing taste and elegant flavor. The National Research Institute of Brewing (NRIB) of Japan has been conducting continuous research and innovations with regard to rice, yeast, A. oryzae, brewing water, production technology, and equipment for sake brewing. The flavor and texture of sake are also a perfect embodiment of the introverted, soft, and medium-characteristic features of Japanese people. This paper reviewed Japanese sake rice, koji, and sake yeast. The characteristics of sake rice, various enzymes in rice koji, and the research advances of sake yeast were expounded in this review.
Various breeding methods and targets for sake yeasts exist, and the genetic engineering of sake yeasts has been attempted.
Eliminating extraneous DNA sequences from the sake yeast genome represents the self-cloning method, and the yeast strains generated with this method do not correspond to genetically modified organisms (GMOs). Therefore, this method will be a promising breeding strategy, which currently awaits public evaluation (Akada, 2002;Akada, Matsuo, Aritomi, & Nishizawa, 1999;Hirosawa et al., 2004;Iijima & Ogata, 2010;Kitagaki & Kitamoto, 2013). In fact, novel methods for the modification of the sake yeast genome without using genetic modification technology that achieves improved brewing performances should be developed.
The palatability of sake paired with different national or region foods could be an interesting and worthy topic for comprehensive exploration, which can further develop sake consumption in the world (https://www.nrib.go.jp/Engli sh/resea rches /resea rch_topi. htm). Certainly, breeding and innovating sake rice cultivars, as well as the selection of novel sake yeast to produce higher-quality sake or sake that meets the specific taste requirements of consumers in particular regions (countries), is a promising way to further promote sake in the world.

ACK N OWLED G M ENTS
The authors gratefully acknowledge support from the Key R&D

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

H U M A N A N D A N I M A L R I G HT S
No animals or humans were used for studies that are the basis of this review.