Reduction of Ethyl Carbamate in an Alcoholic Beverage by CRISPR/Cas9-Based Genome Editing of the Wild Yeast

Ethyl carbamate (EC) is a naturally occurring substance in alcoholic beverages from the reaction of ethanol with urea during fermentation and storage. EC can cause dizziness and vomiting when consumed in small quantities and develop kidney cancer when consumed in excess. Thus, the reduction of EC formation in alcoholic beverages is important for food safety and human health. One of the strategies for reducing EC contents in alcoholic beverages is developing a new yeast starter strain to enable less formation of EC during fermentation. In this study, we isolated a polyploid wild-type yeast Saccharomyces cerevisiae strain from the Nuruk (Korean traditional grain-based inoculum of wild yeast and mold) and developed a starter culture by genome engineering to reduce EC contents in alcoholic beverages. We deleted multiple copies of the target genes involved in the EC formation in the S. cerevisiae by a CRISPR/Cas9-based genome editing tool. First, the CAR1 gene encoding for the arginase enzyme responsible for the formation of urea was completely deleted in the genome of S. cerevisiae. Additionally, the GZF3 gene encoding the transcription factor controlling expression levels of several genes (DUR1, 2, and DUR3) related to urea absorption and degradation was deleted in S. cerevisiae to further reduce the EC formation. The effects of gene deletion were validated by RT-qPCR to confirm changes in transcriptional levels of the EC-related genes. The resulting strain of S. cerevisiae carrying a double deletion of CAR1 and GZF3 genes successfully reduced the EC contents in the fermentation medium without significant changes in alcohol contents and fermentation profiles when compared to the wild-type strain. Finally, we brewed the Korean traditional rice wine Makgeolli using the double deletion strain of S. cerevisiae dCAR1&GZF3, resulting in a significant reduction of the EC content in Makgeolli up to 41.6% when compared to the wild-type strain. This study successfully demonstrated the development of a starter culture to reduce the EC formation in an alcoholic beverage by CRISPR/Cas9 genome editing of the wild yeast.


Introduction
Ethyl carbamate (EC) is a naturally occurring substance during alcoholic fermentation by the reaction of urea with ethanol [1]. EC is known as a carcinogen in experimental animals and possibly carcinogenic to humans (IARC group 2A) [2]. Consumption of EC can cause unconsciousness, vomiting, and damage to the liver and kidneys [1]. As estimated by the JECFA (Joint FAO/WHO Expert Committee on Food Additives), the benchmark dose lower limit of ethyl carbamate is 0.3 mg/kg body weight per day [1]. EC is detected in many alcoholic beverages including beer, wine, brandy, and others at concentrations ranging from 0.01 to 12 mg/L depending on the origin of spirit [1,3]. Considering the high EC contents in alcoholic beverages, several countries have limitations on their levels [1]. For example, maximum levels for ethyl carbamate in wine were limited to 30 µg/L in table wine, 150 µg/L in distilled spirit, and 400 µg/L in fruit brandy in Canada [1]. The reduction of EC concentration in alcoholic beverages is important for ensuring food safety and quality control.
Since EC is constantly produced during fermentation and storage of alcoholic beverages, many efforts have been made to reduce EC contents in fermented foods. First, fermentation temperature can be adjusted to reduce EC formation during yeast fermentation. Second, urease enzyme can be added to the fermentation medium to directly eliminate urea, the major precursor for EC formation. However, urea can be continuously produced from the arginine metabolism of the brewer's yeast Saccharomyces cerevisiae during fermentation. To reduce urea formation, yeast starter strain can be metabolically engineered to modify the arginine metabolism ( Figure 1). Arginine is cleaved into ornithine and urea by arginase encoded by the CAR1 gene in S. cerevisiae. In the previous study, the CAR1 gene was deleted in the S. cerevisiae lab strain, resulting in EC reduction in the growth medium [4]. Additionally, DUR1, 2, and DUR3 genes involved in urea degradation were overexpressed in the Sake yeast to reduce the urea contents [5]. One study attempted to reduce urea and EC in rice wine through modification of expression levels of four GATA transcriptional factors (GLN3, GAT1, DAL80, and GZF3) related to nitrogen catabolite repression in S. cerevisiae [6]. Disruption of the GZF3 transcription factor showed the most effective results for urea and EC reduction in model rice wine [6]. If the industrial strain of S. cerevisiae strain currently used for brewing can be safely engineered for ethyl carbamate reduction, the strain can be directly used for industrial scale production of alcoholic beverages with reduced ethyl carbamate contents.
If the industrial strain of S. cerevisiae strain currently used fo gineered for ethyl carbamate reduction, the strain can be scale production of alcoholic beverages with reduced ethyl c In this study, we aimed to delete both CAR1 and GZF3 charomyces cerevisiae strain isolated from Korean tradition Nuruk containing various wild yeast and molds, and brew a EC contents by the engineered S. cerevisiae (Figure 1). To reli ty issues of GMO products, we engineered the S. cerevisiae nome editing tool which is useful to delete the multiple ge microorganism safely without introducing antibiotics resi Since EC is detected in Makgeolli, a traditional Korean rice w myces cerevisiae [7], we demonstrate the reduction of EC cont with the engineered S. cerevisiae. The EC and ethanol conten the engineered S. cerevisiae and its parental strain were eva deletions on the expression levels of the EC-related genes in tigated.  In this study, we aimed to delete both CAR1 and GZF3 genes in the industrial Saccharomyces cerevisiae strain isolated from Korean traditional fermentation ingredients Nuruk containing various wild yeast and molds, and brew alcoholic beverages with low EC contents by the engineered S. cerevisiae (Figure 1). To relieve the regulatory and safety issues of GMO products, we engineered the S. cerevisiae using the CRISPR/Cas9 genome editing tool which is useful to delete the multiple genes in the genome of target microorganism safely without introducing antibiotics resistant markers or plasmids. Since EC is detected in Makgeolli, a traditional Korean rice wine fermented by Saccharomyces cerevisiae [7], we demonstrate the reduction of EC contents in Makgeolli by brewing with the engineered S. cerevisiae. The EC and ethanol contents in Makgeolli fermented by the engineered S. cerevisiae and its parental strain were evaluated. The impact of gene deletions on the expression levels of the EC-related genes in S. cerevisiae was also investigated.

Materials
The chemical used in the experiment was purchased from Sigma-Aldrich (St. Louis, MO, USA) and media was purchased from Thermo Fisher Scientific (Waltham, MA, USA). The yeast strain was collected from the variable regions of South Korea. Aspergillus oryzae Enzyme Powder ™ was purchased from JEIL BIOTECH Co. (Gyeongju, South Korea) and Japonica rice used from Icheon rice (Icheon, South Korea). All strains and plasmids used in this study are summarized in Table 1. All primers used in this study are summarized in Table 2.

Strains and Plasmids
The Escherichia coli TOP10 strain was used for the construction and propagation of plasmids. The S. cerevisiae GRL6 strain isolated from Nuruk was used for the construction of EC reduction yeast. To isolate yeast strain in Nuruk, 1 g of Nuruk was mixed with 100 mL of DDW and serially diluted. The serially diluted sample was spread onto YPD2 agar (5 g/L yeast extract, 10 g/L peptone, 2 g/L glucose, 1.8 g/L agar) plate and the colonies formed after 2 days of incubation at 30 • C [9]. A single colonization was taken and observed under a microscope to find a group of yeast candidates. The colony PCR was performed using the colony of the candidate group for identification. The ITS1 and ITS4 primers were used to amplify the internal transcribed spacer region of the 18S rRNA gene [10]. The sequence of refined PCR products was sent to BIONICS (Seoul, South Korea) for Sanger sequencing. The BLAST (http://www.ncbi.nlm.nih.gov/BLAST/ (accessed on 18 March 2020) of the National Center for Biotechnology Information (NCBI, MD, USA) was used to obtain homology information [11]. Table 2. Primers used in this study.

Name
Primer Sequence 5 →3 The italic sequences present the homologous region for Gibson assembly. The small letter sequences present the replacement site region for Site Directed Mutagenesis. The underlined sequences present the homologous region for overlap PCR.
The pCas9-Hyg plasmid containing hygromycin B antibiotics marker was constructed to express the Cas9 protein in yeast. The PCR fragment was amplified from the pCas9-NAT plasmid except for Nourseothricin antibiotics marker region by the Cas9-NAT-backbone-F and Cas9-NAT-backbone-R primers. The Hygromycin B-resistant marker was amplified from the pRS42H-HXK2 plasmid by the Hyg-insert-F and Hyg-insert-R primers. These two PCR fragments were assembled by the Gibson Assembly ® Cloning Kit from NEB (Ipswich, MA, USA) to construct the pCas-Hyg plasmid.

Media and Culture Conditions
E. coli was grown in Luria-Bertani (LB) medium (5 g/L yeast extract, 10 g/L tryptone, 10 g/L NaCl) at 37 • C, and ampicillin (50 µg/mL) was added when required. S. cerevisiae was grown in YPD2 medium (5 g/L yeast extract, 10 g/L peptone, 2 g/L glucose) at 30 • C and Nourseothricin sulfate (ClonNAT, 100 µg/mL), Geneticin (G418, 300 µg/mL) were added when required. For comparison of EC formation, batch fermentation was performed in the YPD8 medium (5 g/L yeast extract, 10 g/L peptone, 80 g/L glucose) at pH 6.8 and 30 • C. Arginine 50 mM and urea 50 mg/L which are similar levels to the raw ingredients of alcoholic beverages were added to the medium to provide the EC formation condition [12,13].

Chromosomal Gene Deletion by CRISPR/Cas9 System
After the DNA double-strand break using the CRISPR/Cas9 system, the S. cerevisiae GRL6 strain was repaired by the 90 bp donor DNA through homologous recombination. The donor DNA of CAR1 gene deletion was constructed by the Donor-CAR1-F and Donor-CAR1-R primers. The donor DNA of GZF3 gene deletion was constructed by the Donor-GZF3-F and Donor-GZF3-R primers. Each 60 bp forward and reverse primer contains a complementary 30 bp homologous region.
The pCas9-Hyg, pRS42K-gRNA-CAR1, and the CAR1 donor DNA were transformed into S. cerevisiae GRL6 for the CAR1 gene deletion. The pCas9-Hyg, pRS42K-gRNA-GZF3, and the GZF3 donor DNA were transformed into S. cerevisiae GRL6 and S. cerevisiae ∆CAR1 for GZF3 gene deletion, respectively. The transformation was performed by the highefficiency yeast transformation method using the LiAc/SS carrier DNA/PEG [14]. After the transformation, the gene deletion was confirmed by the colony PCR. The CAR1 gene deletion was checked by the Check-CAR1-F and Check-CAR1-R primers (WT: 3112 bp, CAR1 deletion: 2100 bp) and the GZF3 gene deletion was checked by the Check-GZF3-F and Check-GZF3-R primers (WT: 2756 bp, GZF3 deletion: 900 bp).

Flow Cytometry Analysis of Yeast Ploidy
The method referred to the Propidium Iodide Nucleic Acid Stain Manuals by Invitrogen (Waltham, MA, USA) and previous ploidy analysis studies [15,16]. The yeast cells were inoculated into 5 mL of YPD2 medium (20 g/L glucose) at 30 • C. When the optical density of the culture reached to 1.0 (OD 600 ), the cells were collected and resuspended in 500 µL of water and centrifuged for 3 min (500× g). After removing the supernatant, 1 mL of phosphate-buffered solution (PBS) at room temperature and 4 mL absolute ethanol at −20 • C were added to the cell pellet and slowly resuspended. After vortexing of the cell suspension, the cells stayed in ethanol at −20 • C for 20 min. Then, the solution was centrifuged for 1 min (3000× g), and the cells were harvested and rehydrated by adding 5 mL of PBS for 15 min. Rehydrated cells were centrifuged at 3 min (3000× g) and the supernatant was removed. One mL of 3 µM Propidium Iodide (PI) solution in buffer (100 mM Tris, pH 7.4, 1 mM CaCl 2 , 150 mM NaCl, 0.5 mM MgCl 2 , 0.1% Nonidet P-40) was added and slowly resuspended. The samples stayed for 1 h at room temperature in the dark. The stained cells were stored at 4 • C until analysis. The samples were briefly sonicated and analyzed with Cytek ® Biosciences Flow Cytometry Analyzer (NL-1000) using a Texas Red filter. Histograms were acquired in linear mode. The data were analyzed with SpectroFlo ® software (Cytek Biosciences, Fremont, CA, USA).

RNA Preparation and Real-Time Quantitative PCR Assay
We collected yeast cells grown for 6 h in batch fermentation and used them for RNA preparation and Real-Time Quantitative PCR (RT-qPCR) Assay. The preparation of yeast RNA was performed by AccuPrep ® Universal RNA Extraction Kit from BIONEER (Daejeon, Korea). The detailed experimental methods followed the AccuPrep ® Universal RNA Extraction Kit Protocol. The cDNA synthesis of yeast RNA was carried out using the PrimeScript ™ 1st strand cDNA Synthesis Kit from Takara Bio (Shiga, Japan) and its protocol.
The RT-qPCR was performed to determine the RNA expression using a Real-Time PCR instrument (Applied Biosystems, CA, USA). The parameters for PCR were as follows: pre-incubation at 95 • C for 30 s, then 40-50 cycles of amplification at 95 • C for 5 s, 55 • C for 20 s, and cooling at 50 • C for 30 s. The relative expression of the CAR1, DUR1, 2, and DUR3 genes was determined by the ∆∆Ct method based on the ACT1 gene which is known to be continuously expressed [17,18]. Primers used for RT-qPCR analysis are summarized in Table 2.

Brewing Korean Rice Wine Makgeolli
The Korean rice wine Makgeolli was brewed to measure the EC production during the fermentation by the engineered yeasts and its parent strain. The 100 g of japonica rice was steamed for 30 min and cooled. The yeast was cultivated in YPD2 medium overnight at 30 • C. For brewing of Makgeolli, 300 mL of water, steamed rice, 10 g of Aspergillus oryzae enzyme powder, and yeast cells OD 600 = 50 (around 15 g DCW, and 0.7 × 10 9 cells) were mixed. Fermentation was performed for 10 days at 20 • C by stationary culture.

Fermentation Capability and EC Analysis
The optical density of yeast culture was measured at 600 nm absorbance using a UV-1800 spectrophotometer (SHIMADZU, Kyoto, Japan). For analysis of fermented products, culture supernatants were diluted appropriately after centrifugation of the batch fermentation medium and Makgeolli. Concentrations of glucose, glycerol, acetate, and ethanol were measured using the 1260 Infinity II HPLC system of Agilent Technologies (Santa Clara, CA, USA) equipped with an Aminex ® HPX-87H column (300 × 7.8 mm) (Bio-Rad, Hercules, CA, USA). The mobile phase was 5 mM H 2 SO 4 solution and the flow rate was 0.6 mL/min. The column was heated at 60 • C to analyze 10 µL of a diluted supernatant sample using the previous method [4].
For EC quantification, standards and samples were prepared according to a previous study [19]. The analysis of EC content was performed using the 6890 N gas chromatograph (Agilent Technologies, CA, USA) equipped with the 5973 Mass Selective Detector (GC-MS) (Agilent Technologies, CA, USA). The samples were separated on a DB-5 MS column (30 m × 0.250 mm ID × 0.25 µm) (Agilent Technologies, CA, USA). The flow rate of the helium carrier gas was 0.5 mL/min, and 1 µL of the extracts was injected with splitless mode. The injector temperature was 200 • C and the oven was programmed as following conditions (50 • C held for 2 min, 2 • C/min increased to 100 • C, and 30 • C/min increased to 200 • C, held for 10 min). The mass spectral identification of compounds was applied at an ion source temperature of 230 • C and mass ionization voltage of 70 eV. The selected ion monitoring (SIM: m/z 62, 74, 89) was used for the analysis of EC. For quantitative analysis of EC, m/z 62 which is the major fragment ion was used and has few interruptions between fragments [20].

Statistical Analysis
At least three repeat experiments were performed for each fermentation experiment. The results were expressed as average with standard errors. All statistical analyses were performed using SAS 9.4 from SAS Institute Inc. (Cary, NC, USA). Data were analyzed using One-way ANOVA. p-values of less than 0.05 were considered statistically significant.

Isolation of Yeast Stater Strain and Ploidy Analysis
To develop a starter culture to reduce EC contents in alcoholic beverages, we first isolated the wild-type Saccharomyces cerevisiae strain from the Nuruk, Korean traditional grain-based inoculum of wild yeast and mold. The Nuruk sample was diluted with water and spread onto the YPD2 agar plate. After 2 days of incubation, several hundred colonies were grown on the plate. Among them, a single colony was selected based on ethanol production phenotype. The isolated colony named GRL6 was observed under a microscope and showed a yeast-like shape. For identification of the isolated strain, the colony PCR of the GRL6 strain was performed to check the internal transcribed spacer region of the 18S rRNA gene. Sanger sequencing and BLAST analysis suggested that the isolated strain GRL6 was identified as S. cerevisiae.
To confirm the ploidy of the isolated wild-type S. cerevisiae GRL6 from Nuruk, we performed a ploidy analysis of the cells stained with PI by flow cytometry. A well-known laboratory haploid strain, S. cerevisiae BY4742 which is a well-known laboratory haploid strain was used to compare ploidy [21,22]. Based on the analysis, S. cerevisiae GRL6 was estimated as a diploid strain when compared to the haploid S. cerevisiae BY4742 (Figure 2). This result suggested that the isolated strain of S. cerevisiae GRL6 has multiple copies of genes in the chromosome of the cells. Since the CRISPR/Cas9 system can be utilized for the simultaneous deletion of multi-copy genes in the chromosomes, we decided to delete several genes related to the EC formation by the Cas9 genome editing tool.

Deletion of CAR1 and GZF3 in Industrial Yeast Strain using CRISPR/Cas9 System
To reduce EC production in alcoholic beverages by developing a new starter culture, we attempted to delete the CAR1 and GZF3 genes in the newly isolated wild-type yeast strain, S. cerevisiae GRL6. Since the CAR1 gene encodes the arginase enzyme responsible for the formation of urea in S. cerevisiae, the CAR1 gene can be deleted for EC reduction. Additionally, the GZF3 gene encoding the transcription factor controlling expression levels of the DUR1, 2 (urea amidolyase) and DUR3 (urea active transporter) genes related to urea absorption and degradation can be deleted in S. cerevisiae for EC reduction phenotype [23,24] ( Figure 1). We used the CRISPR/Cas9 genome editing tool to carry out the double deletion of CAR1 and GZF3 genes in the diploid S. cerevisiae GRL6 ( Figure S1).
The Cas9-Hyg plasmid was constructed and used for the Cas9 expression for targetspecific double-strand DNA breakage in the genome of S. cerevisiae. The pRS42K-gRNA-CAR1 and pRS42K-gRNA-GZF3 plasmids were constructed for the sgRNA expression to guide the CAR1 and GZF3 gene deletion by Cas9, respectively. Through the transformation of Cas9 and sgRNA expression plasmids and donor DNA for gene deletion, multi-copies of CAR1 and GZF3 genes were deleted in the diploid S. cerevisiae GRL6. As a result, we constructed the engineered S. cerevisiae dCAR1 strain that deleted only CAR1 genes, the dGZF3 strain that deleted only GZF3 genes, and the dCAR1&GZF3 strain that deleted both CAR1 and GZF3 genes. Deletions of the CAR1 and GZF3 genes in the engineered S. cerevisiae strains were confirmed by colony PCR ( Figure S2). The result suggests that complete deletion of the CAR1 and GZF3 genes in multiple chromosomes was successfully carried out in the wild-type yeast S. cerevisiae GRL6 strain isolated from Nuruk.

Real-Time Quantitative PCR Analysis of the Engineered S. cerevisiae
The engineered strain of S. cerevisiae carrying the double deletion of CAR1 and GZF3 genes was successfully constructed by the CRISPR/Cas9 tool in this study. To confirm the effects of gene deletion, we performed the RT-qPCR analysis and checked changes in transcriptional levels of the EC-related genes. The previous study carrying out the GZF3 gene deletion in S. cerevisiae confirmed the increased RNA expression levels of DUR1, 2, and DUR3 genes [6]. Another study carrying out the CAR1 gene deletion in S. cerevisiae confirmed no expression of the CAR1 through RT-qPCR [18]. Since we constructed the engineered S. cerevisiae strain harboring the double deletion of CAR1 and GZF3 genes for the first time, we performed the RT-qPCR analysis to confirm the changes in RNA expression levels of the CAR1, DUR1, 2 and DUR3 genes in the engineered strains. The relative RNA expressions of the S. cerevisiae dCAR1, dGZF3, and dCAR1&GZF3 strains suggested the significant differences in the expression levels of the CAR1, DUR1, 2, and DUR3 genes (Figure 3).
The expression of the CAR1 gene was not detected in the engineered dCAR1 and dCAR1&GZF3 because the CAR1 gene was completely deleted in these strains. Meanwhile, the engineered S. cerevisiae dGZF3 strain exhibited the expression of the CAR1 gene with no significant difference when compared to that of the wild-type S. cerevisiae GRL6. For the DUR1, 2 gene expression, the dCAR1 strain showed no significant difference with the wildtype GRL6. However, the dGZF3 and dCAR1&GZF3 strains harboring the complete deletion of the GZF3 gene showed a significant difference in the DUR1 and DUR2 gene expression levels when compared to the wild-type GRL6. The DUR1 and DUR2 gene expression levels increased by 2.12 times and 1.92 times in the dGZF3 and the dCAR1&GZF3 strains, respectively. For the DUR3 gene expression, the dCAR1 strain carrying only CAR1 gene deletion showed no significant difference compared to the wild-type strain. The dGZF3 and dCAR1&GZF3 strains exhibited increased DUR3 gene expression levels by 1.41 times and 1.31 times, respectively, when compared to the wild-type S. cerevisiae GRL6. These results indicate that the effects of gene deletion in the engineered S. cerevisiae strains were successfully validated as we expected based on the previous studies. The engineered S. cerevisiae carrying the double deletion of CAR1 and GZF3 genes showed no expression of the CAR1 and upregulation of DUR1, 2, and DUR 3 expression, which may help to reduce EC formation during fermentation.
GZF3 genes for the first time, we performed the RT-qPCR analysis to c es in RNA expression levels of the CAR1, DUR1, 2 and DUR3 genes strains. The relative RNA expressions of the S. cerevisiae dCA dCAR1&GZF3 strains suggested the significant differences in the expre CAR1, DUR1, 2, and DUR3 genes (Figure 3). Figure 3. Determination of CAR1, DUR1, 2 and DUR3 gene expression levels cerevisiae GRL6, and the engineered S. cerevisiae dCAR1, dGZF3, and dCAR1& in the YPD8 medium with 50 mg/L urea and 50 mM arginine at 30 °C for 6 h. used as a reference for RT-qPCR analysis using the ΔΔCt method. Data were a way ANOVA to determine whether the expression levels of each gene in four tistically significant differences (a and b) between the means of the groups. than 0.05 were considered statistically significant. N.D.: Not Detected.
The expression of the CAR1 gene was not detected in the engin dCAR1&GZF3 because the CAR1 gene was completely deleted in the while, the engineered S. cerevisiae dGZF3 strain exhibited the expres gene with no significant difference when compared to that of the wild GRL6. For the DUR1, 2 gene expression, the dCAR1 strain showed no ence with the wild-type GRL6. However, the dGZF3 and dCAR1&GZ ing the complete deletion of the GZF3 gene showed a significant differ and DUR2 gene expression levels when compared to the wild-type and DUR2 gene expression levels increased by 2.12 times and 1.92 ti and the dCAR1&GZF3 strains, respectively. For the DUR3 gene expre strain carrying only CAR1 gene deletion showed no significant differ the wild-type strain. The dGZF3 and dCAR1&GZF3 strains exhibited gene expression levels by 1.41 times and 1.31 times, respectively, when wild-type S. cerevisiae GRL6. These results indicate that the effects of ge engineered S. cerevisiae strains were successfully validated as we expe previous studies. The engineered S. cerevisiae carrying the double dele GZF3 genes showed no expression of the CAR1 and upregulation of D Figure 3. Determination of CAR1, DUR1, 2 and DUR3 gene expression levels in the wild-type S. cerevisiae GRL6, and the engineered S. cerevisiae dCAR1, dGZF3, and dCAR1&GZF3 strains grown in the YPD8 medium with 50 mg/L urea and 50 mM arginine at 30 • C for 6 h. The ACT1 gene was used as a reference for RT-qPCR analysis using the ∆∆Ct method. Data were analyzed using One-way ANOVA to determine whether the expression levels of each gene in four strains are any statistically significant differences (a and b) between the means of the groups. The p-values of less than 0.05 were considered statistically significant. N.D.: Not Detected.

EC Reduction in Fermentation Medium by the Engineered S. cerevisiae
To evaluate the EC reduction and fermentation capability, batch fermentations in YPD8 medium containing 80 g/L glucose with 50 mg/L urea and 50 mM arginine were conducted by the engineered S. cerevisiae dCAR1, dGZF3, and dCAR1&GZF3 strains and its parent strain GRL6 ( Figure S3). The engineered S. cerevisiae strains did not show significant differences in fermentation profiles when compared to the wild-type GRL6. Final OD 600 value, ethanol titer, yield, and productivity after 12 h fermentation of the engineered S. cerevisiae strains were similar to those of the wild-type strain (Table S1). These results indicated that the CAR1 and GZF3 gene deletion did not affect the fermentation capability of S. cerevisiae GRL6.
Meanwhile, the EC concentrations in the batch fermentation by the engineered strains were significantly different than those of the wild-type S. cerevisiae GRL6 (Figure 4). At 24 h of batch fermentation, the EC formation decreased by 32.1% and 16.9% in the dCAR1 and dGZF3 strains, respectively. The EC concentration by the dCAR1&GZF3 strain decreased up to 52.1% when compared to the wild-type GRL6. At 168 h of batch fermentation, the EC concentration decreased by 26.1% and 15.3% in the dCAR1 and dGZF3 strains, respectively. The dCAR1&GZF3 strain harboring a double deletion of the CAR1 and GZF3 genes showed 33.9% of EC reduction when compared to the wild-type GRL6 (Table S2). These results suggested that deletion of both CAR1 and GZF3 genes in S. cerevisiae showed the highest reduction in EC concentration during batch fermentation in yeast growth medium containing urea and arginine. CAR1 and GZF3 genes showed 33.9% of EC reduction when compared to GRL6 (Table S2). These results suggested that deletion of both CAR1 and G S. cerevisiae showed the highest reduction in EC concentration during batch in yeast growth medium containing urea and arginine.

EC Reduction in Makgeolli Brewed by the Engineered S. cerevisiae
To evaluate the EC reduction in alcoholic beverages brewed by th strains, we brewed Makgeolli, the Korean traditional rice wine using the cerevisiae dCAR1, dGZF3, and dCAR1&GZF3 strains and its parent strain G cultures ( Figure S4). There was no significant difference in the ethanol co the Makgeolli products brewed by the engineered S. cerevisiae strains when

EC Reduction in Makgeolli Brewed by the Engineered S. cerevisiae
To evaluate the EC reduction in alcoholic beverages brewed by the engineered strains, we brewed Makgeolli, the Korean traditional rice wine using the engineered S. cerevisiae dCAR1, dGZF3, and dCAR1&GZF3 strains and its parent strain GRL6 as starter cultures ( Figure S4). There was no significant difference in the ethanol concentration of the Makgeolli products brewed by the engineered S. cerevisiae strains when compared to the wildtype GRL6 ( Figure S4). The EC concentrations in Makgeolli brewed by the dCAR1 and dGZF3 strains slightly decreased when compared to the wild-type strain at both 24 h and 168 h ( Figure 5). However, the dCAR1&GZF3 strain significantly decreased the EC concentration in Makgeolli up to 37.0% when compared to the wild-type GRL6 strain after 24 h of the brewing (Table S3). After 168 h, the EC concentration in Makgeolli brewed by the dCAR1&GZF3 strain decreased by 41.6%. These results suggested that the deletion of CAR1 and GZF3 genes in S. cerevisiae helped to decrease the EC concentration in alcoholic beverages. Especially, the double deletion of CAR1 and GZF3 genes was most effective to reduce the EC concentration among the engineered S. cerevisiae strains.
ed that the deletion of CAR1 and GZF3 genes in S. cerevisiae helped to de concentration in alcoholic beverages. Especially, the double deletion of CA genes was most effective to reduce the EC concentration among the engin visiae strains.

Conclusions
In this study, we successfully isolated the wild-type S. cerevisiae strain the Korean traditional Fermentation ingredients used for brewing the Kore rice wine Makgeolli. The isolated strain was identified as a diploid S. cerev ploidy analysis using flow cytometry and named the wild-type S. cerevis develop this strain as a starter culture strain reducing EC formation in alc ages, the CRISPR/Cas9 genome engineering tool was utilized for multiple g in the yeast genome. The CAR1 and GZF3 genes were selected for gene d on previous studies and successfully deleted in the genome of the S. ce strain by the Cas9 for the first time. The effect of CAR1 and GZF3 double cerevisiae was validated by the RT-qPCR analysis which confirmed upreg DUR1, 2, and DUR3 expression without the CAR1 expression. Batch ferme engineered S. cerevisiae dCAR1, dGZF3, and dCAR1&GZF3 in yeast growth taining urea and arginine showed a significant reduction of EC concentrat to the wild-type S. cerevisiae GRL6 without affecting fermentation capabil The double deletion strain dCAR1&GZF3 decreased the most by 52.1%. T tion in Makgeolli was also demonstrated by brewing with the engineered S. double deletion strain dCAR1&GZF3 significantly decreased the EC concen 41.6% after 168 h of fermentation without affecting ethanol concentratio Figure 5. Comparison of EC concentrations in Makgeolli products brewed by the wild-type S. cerevisiae GRL6, and the engineered S. cerevisiae dCAR1, dGZF3, and dCAR1&GZF3 strains. Data were analyzed using One-way ANOVA to determine whether the EC concentrations in Makgeolli by the engineered strains are any statistically significant differences (a and b) when compared to the wild-type strain at each time point (24 h or 168 h). * means p < 0.05.

Conclusions
In this study, we successfully isolated the wild-type S. cerevisiae strain from Nuruk, the Korean traditional Fermentation ingredients used for brewing the Korean traditional rice wine Makgeolli. The isolated strain was identified as a diploid S. cerevisiae strain by ploidy analysis using flow cytometry and named the wild-type S. cerevisiae GRL6. To develop this strain as a starter culture strain reducing EC formation in alcoholic beverages, the CRISPR/Cas9 genome engineering tool was utilized for multiple gene deletions in the yeast genome. The CAR1 and GZF3 genes were selected for gene deletion based on previous studies and successfully deleted in the genome of the S. cerevisiae GRL6 strain by the Cas9 for the first time. The effect of CAR1 and GZF3 double deletion in S. cerevisiae was validated by the RT-qPCR analysis which confirmed upregulation of the DUR1, 2, and DUR3 expression without the CAR1 expression. Batch fermentation of the engineered S. cerevisiae dCAR1, dGZF3, and dCAR1&GZF3 in yeast growth medium containing urea and arginine showed a significant reduction of EC concentration compared to the wild-type S. cerevisiae GRL6 without affecting fermentation capability of strains. The double deletion strain dCAR1&GZF3 decreased the most by 52.1%. The EC reduction in Makgeolli was also demonstrated by brewing with the engineered S. cerevisiae. The double deletion strain dCAR1&GZF3 significantly decreased the EC concentration up to 41.6% after 168 h of fermentation without affecting ethanol concentration. This study successfully demonstrated the method of reducing EC concentration in alcoholic beverages by genome engineering of the wild S. cerevisiae starter culture using the CRISPR/Cas9 system. The engineered S. cerevisiae strains constructed in this study can be applied to reduce the EC concentration in other alcoholic beverages such as fruits wine.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/foods12010102/s1, Figure S1: Diagram of the CAR1 and GZF3 gene deletion using CRISPR/Cas9 system; Figure S2: Agarose gel electrophoresis (1% agarose) of colony PCR products; Figure S3: Batch fermentation profiles of S. cerevisiae GRL6, dCAR1, dGZF3, and dCAR1&GZF3 in the YPD8 medium containing 80 g/L glucose with 50 mg/L urea and 50 mM arginine for 24 h; Figure S4: Brewing alcoholic beverages by using the engineered S. cerevisiae as a starter culture; Table S1: Comparison of the fermentation capability in batch fermentation at 12 h; Table S2: Comparison of the ethyl carbamate concentration in batch fermentation; Table S3: Comparison of the ethyl carbamate concentration in Makgeolli fermentation by the wild-type S. cerevisiae GRL6 and the engineered S. cerevisiae dCAR1, dGZF3, and dCAR1&GZF3.