Characterization of the Psychrotrophic Lactic Acid Bacterium Leuconostoc gelidum subsp. aenigmaticum LS4 Isolated from Kimchi Based on Comparative Analyses of Its Genomic and Phenotypic Properties

With the aim of developing a new food starter culture, twenty-three psychrotrophic lactic acid bacteria (LAB) were isolated from 16 kimchi samples. One strain, Leuconostoc gelidum subsp. aenigmaticum LS4, which had typical psychrotrophic characteristics, was selected, and its phenotypic and genomic properties as a starter culture were investigated. The complete genome of L. aenigmaticum LS4 consisted of one circular chromosome (1,988,425 bp) and two plasmids (19,308 bp and 11,283 bp), with a guanine–cytosine content of 36.8%. L. aenigmaticum LS4 could grow at 5 °C but not at 37 °C, and maximum cell growth was obtained at 15~25 °C. L. aenigmaticum LS4 did not show any harmful characteristics such as hemolysis, undesirable enzyme activities, biogenic amine production, or antibiotic resistance. L. aenigmaticum LS4 was investigated for its suitability for technological processes (pH, temperature and NaCl treatment). L. aenigmaticum LS4 exhibited strong antimicrobial activity caused by the production of organic acids and bacteriocin, and it produced an exopolysaccharide composed of glucose with a molecular weight of 3.7 × 106 Da. Furthermore, L. aenigmaticum LS4 improved the organoleptic qualities of kimchi juice. Our results indicate that L. aenigmaticum LS4 could be used as a functional starter culture for food (vegetable or fruit) fermentation at low temperatures.


Introduction
Kimchi, a traditional Korean fermented vegetable, is well known for its health benefits, which include anticancer, anti-obesity, immune promotion, brain health promotion, cholesterol reduction, and antioxidative effects [1,2]. Numerous lactic acid bacteria (LAB) are involved in kimchi fermentation; thus, kimchi is considered a good source of potentially beneficial LAB [3]. The growing consumer demand for healthier foods is stimulating new product development in the food industry worldwide. Various LAB strains were isolated from kimchi with the aim of developing new probiotics or starter cultures for foods [3], but most of these isolates were mesophilic LAB due to the use of a 30 • C incubation temperature in isolation experiments [4][5][6]. However, kimchi is generally fermented at temperatures of less than 10 • C [7]; thus, it can be inferred that LAB that can grow well at low temperatures play an important role in kimchi fermentation.
Active LAB populations responsible for kimchi fermentation are of the genera Lactobacillus, Leuconostoc, or Weissella as determined by culture-independent and culturedependent methods [8]. Thus far, LAB strains that grow well at 30 • C have been isolated, such as Leuconostoc mesenteroides, Leuconostoc citreum, Leuconostoc kimchii, Lactobacillus plantarum, Lactobacillus sakei, Weissella confusa, Weissella cibaria, and Weissella koreensis [3,[9][10][11]. On the other hand, psychrotrophic LAB, which barely grow at 30 • C, have Carbohydrate utilization by the LAB isolates was determined using API 50 CHL (BioMérieux, Marcy l'Etoile, France), according to the manufacturer's instructions. Isolates were inoculated into the API strip and incubated for 48 h at 25 • C, and acid production from the supplied carbohydrate was determined as described by the manufacturer.

Phenotypic Characteristics as Starter Cultures 2.4.1. Sensory Properties
We evaluated the effects of the LAB isolates on the organoleptic qualities of kimchi juice. To prepare starter cultures, LAB isolates cultivated overnight in MRS were centrifuged at 9950× g for 15 min at 4 • C, washed with sterile distilled water, and resuspended in the same volumes of distilled water. To prepare kimchi juice, freshly prepared nonfermented kimchi was macerated using a juice maker (HD-RBF09, Hurom, Seoul, Korea), stored at −20 • C, and thawed under running tap water when required. The prepared starter cultures were inoculated at 5.0% (v/v) into prepared kimchi juice and fermented for 24 h at 25 • C; thereafter, its sensory characteristics were investigated. The kimchi juice without starter inoculum was used as a control.
Sensory evaluations were carried out after obtaining approval from the Institutional Review Board of Chosun University (IRB # 2-1041055-AB-N-01-2020-29). Nine trained graduate students, who had performed more than 30 sensory evaluations on kimchi per year at the Kimchi Research Center, Chosun University (Gwangju, Korea), conducted the evaluation. Samples (10 mL) labeled with a random 3-digit code were served on white bowls. The panelists evaluated sensory attributes using 5-point scales for sourness, fresh taste, carbonate taste, pleasant fermentative smell, sewerage-like smell (1 = very weak, 3 = moderate, and 5 = very strong), and overall acceptance (1 = very bad, 3 = moderate, and 5 = very good).

Hemolysis and Antibiotic Resistance
Hemolysis was examined by streaking LAB cells on blood agar containing 7.0% horse blood (Oxoid, Hampshire, UK). Plates were incubated for 48 h at 25 • C to detect α-hemolysis or for 48 h at 25 • C and then kept at 4 • C for 24 h to detect β-hemolysis. The clear zones around colonies were then observed [9,18].
Antibiotic susceptibilities of LAB isolates were evaluated according to the technical guidelines issued by the European Food Safety Authority (EFSA) [19]. Minimal inhibitory concentrations (MICs) of ampicillin, chloramphenicol, erythromycin, gentamycin, kanamycin, streptomycin, tetracycline, and vancomycin (Sigma) were determined [9]. Briefly, LAB isolates cultivated overnight were centrifuged (9950× g, 15 min, 4 • C) and resuspended in Mueller-Hilton (MH; Difco) broth supplemented with 0.5% dextrose (final cell concentration; 7.0 log CFU/mL). Antibiotics were then added to aliquots of MH suspension and incubated anaerobically for 24~48 h at 25 • C [9]. Growth of LAB isolates was determined at 600 nm (Biochrom). MH cultures not treated with antibiotics were used as controls.

Enzymatic Activities
API-ZYM kits (BioMérieux) were used to determine the enzymatic activities of LAB isolates. LAB were cultivated, harvested, and resuspended (McFarland standard 5) in sterile distilled water, and suspensions were then spotted (65 µL) into wells and incubated for 4 h at 25 • C. Then, one drop of the kit reagents ZYM-A and ZYM-B was then added to each well, and after allowing reactions to proceed for 5 min, enzymatic activity was read [3].

Stress Tolerance: Temperature, pH, and NaCl concentration
To determine the effect of temperature on cell viability, the selected LAB isolate was cultivated in MRS broth for 24 h at 25 • C, harvested, resuspended in MRS broth of the same volume (final cell concentration; 9.40 log CFU/mL), and then incubated for 72 h at −2, 0, 4, 10, 20, 30, 50, or 70 • C.
Cell viability of LAB isolate under acid and alkali conditions was also investigated. LAB isolate was cultivated in MRS broth for 24 h at 25 • C, harvested, resuspended in MRS broth adjusted to pH 2.0~10.0 of the same volume, and incubated for 72 h at 25 • C.
To determine the salt tolerance of LAB isolate, LAB isolate cultivated for 24 h was harvested, resuspended in MRS broth containing 1~15% NaCl of the same volume, and incubated for 72 h at 25 • C.
During the 72 h incubation periods of the temperature, pH, and NaCl concentration tolerance experiments, viable cell counts (CFU/mL) were determined every 24 h using a plate counting method [3].
To purify LAB-produced EPS, the selected LAB isolate was cultured for 48 h at 25 • C in sucrose broth medium and the EPS produced was purified by ethanol precipitation [23]. Briefly, culture was centrifuged (9950× g, 25 min, 4 • C), trichloroacetic acid (4%; v/v) was added, and then denatured proteins were removed by centrifugation. Supernatant was passed through a 0.4 µm filter (Advantec, Tokyo, Japan) and mixed with two volumes of 95% ethanol for 16 h at 4 • C. EPS was harvested, resuspended in distilled water, and dialyzed (MW cut-off 10,000 Da; Spectrum, Rancho Dominguez, CA, USA) for 24 h at 4 • C. The dialyzed solution obtained was freeze-dried and used for further analysis.
To identify the monosaccharide composition of EPS, purified EPS was hydrolyzed (5 N H 2 SO 4 , 6 h, 100 • C) and analyzed by HPLC (Ultimate 3000, Thermo Scientific Dionex, Sunnyvale, CA, USA) using an Aminex 87H column (300 × 10 mm; BioRad) and an RI detector (Refracto MAX520, Tokyo, Japan). H 2 SO 4 (0.01 N) was used as the elution buffer at a flow rate of 0.5 mL/min. Gel permeation chromatography (GPC) was performed using a serial set of three columns (Ultrahydrogel 120, 500, and 1000; Waters, Milford, MA, USA). Sodium azide in water (0.1 M) was used as an elution buffer at a flow rate of 1 mL/min. The data obtained were used to determine the molecular weight of the purified EPS using a software package (Chromeleon ver. 6.8, Dionex, CA, USA).

Antimicrobial Activity
Antimicrobial activities of the selected LAB isolate against pathogenic bacteria and food spoilage molds were assayed using the paper disc assay [24]. The microbial strains used and their culture media and culture conditions are listed in Supplementary Table S1. ATCC strains were purchased from the American Type Culture Collection (Manassas, VA, USA), KCCM strains from the Korean Culture Center of Microorganisms (Seoul, Korea), and PF strains were isolated in our laboratory [25].
To prepare culture filtrate of the LAB isolate for antimicrobial studies, LAB was cultivated in MRS broth for 48 h at 25 • C, harvested, and filtered (0.4 µm pore size, Advantec). The resulting filtrate was used as an antimicrobial sample. Simultaneously, the culture filtrate treated with protease (2 mg/mL; Sigma) for 6 h at 37 • C was also used as an antimicrobial sample. Moreover, organic acids produced by LAB isolate was determined by HPLC using the method used for identifying monosaccharides in EPS (Section 2.4.6), and then the organic acid mixture in the culture filtrate quantified was also used as an antimicrobial sample.
Antimicrobial assay and evaluation of antibacterial and antifungal activities were performed using 1X and 5X concentrated antimicrobial samples, respectively. For antibacterial assays, plates were prepared by spreading each pathogen at 6 log CFU/mL on an appropriate medium, and then paper discs (diameter 8 mm; Advantec) were placed on the medium and 100 µL aliquots of antibacterial samples were spotted onto paper discs. For antifungal assays, plates were prepared by adding mold (6 log spores/20 mL of MEA or PDA) to 1.5% (w/v) Bacto agar (Duchefa, Harlem, The Netherlands) as listed in Supplementary Table S1. Spore solutions were prepared as previously described [24]. Paper discs (diameter 8 mm; Advantec) on MEA or PDA plates were spotted (100 µL) with the prepared antifungal samples. Plates were incubated for 24~48 h at 25~37 • C, and the diameters of inhibition zones around colonies were measured using a caliper (CD-15CPX, Mitutoyo, Kawasaki, Japan).

Genome Sequencing and Analysis
Genomic DNA extraction was performed using DNeasy Blood & Tissue Kits (Qiagen), and whole-genome sequencing of the selected LAB isolate was performed as previously described [7]. The complete genome of LAB isolate was sequenced by Illumina HiseqXten sequencing and PacBio RS single-molecule real-time sequencing and constructed de novo using the hierarchical genome assembly process; paired-end reads were obtained by Illumina sequencing by Macrogen (Seoul, Korea). The genome sequences of the chromosome and two plasmids have been deposited in GenBank (accession numbers: CP071950~2).
Whole-genome sequences were annotated via the RAST server using the RASTtk scheme (https://rast.nmpdr.org/ (accessed on 10 January 2021)) and the KEGG database using BlastKOALA web tool (http://www.kegg.jp/blastkoala/ (accessed on 12 May 2021)). The general and functional genome features of the LAB isolate were analyzed using RAST annotation and BlastKOALA results. Putative bacteriocin genes were identified using BAGEL 4. Antibiotic resistance genes were analyzed using ResFinder ver. 4.1.

Statistical Analysis
Results were statistically analyzed using the Statistical Package for the Social Science (SPSS) program (Version 26.0 for Window, Chicago, IL, USA) and data were expressed as means ± standard deviations (SD). All experiments were conducted three times with duplicate determinations. Means differences among data were assessed by Duncan's multiple range test (p < 0.05).

Microorganisms Present in Kimchi Samples, and Isolation and Identification of Psychrotropic LAB
Microbial analysis was performed on kimchi samples stored at 0~10 • C for 1~4 months by PCR-DGGE ( Figure 1). The microbial profiles of the 16 kimchi samples were slightly different; however, three bands (a~c; arrowed) were detected as main bands in most of the samples. Bands a, b, and c were identified as Lb. sakei (with 99.8% identity), W. koreensis (with 100% identity), and Leuconostoc gelidum or Leuconostoc inhae (with 100% identities), respectively. The same results were obtained for kimchi fermented at −1.5~0 • C for 2~3 months in our previous study [7]. Kim et al. also reported that L. gelidum is a dominant Leuconostoc species in kimchi fermented at 8 • C [26]. obtained by Illumina sequencing by Macrogen (Seoul, Korea). The genome sequences of the chromosome and two plasmids have been deposited in GenBank (accession numbers: CP071950~2).
Whole-genome sequences were annotated via the RAST server using the RASTtk scheme (https://rast.nmpdr.org/ (accessed on 10 January 2021)) and the KEGG database using BlastKOALA web tool (http://www.kegg. jp/blastkoala/ (accessed on 12 May 2021)). The general and functional genome features of the LAB isolate were analyzed using RAST annotation and BlastKOALA results. Putative bacteriocin genes were identified using BA-GEL 4. Antibiotic resistance genes were analyzed using ResFinder ver. 4.1.

Statistical Analysis
Results were statistically analyzed using the Statistical Package for the Social Science (SPSS) program (Version 26.0 for Window, Chicago, IL, USA) and data were expressed as means ± standard deviations (SD). All experiments were conducted three times with duplicate determinations. Means differences among data were assessed by Duncan's multiple range test (p < 0.05).

Microorganisms Present in Kimchi Samples, and Isolation and Identification of Psychrotropic LAB
Microbial analysis was performed on kimchi samples stored at 0~10 °C for 1~4 months by PCR-DGGE ( Figure 1). The microbial profiles of the 16 kimchi samples were slightly different; however, three bands (a~c; arrowed) were detected as main bands in most of the samples. Bands a, b, and c were identified as Lb. sakei (with 99.8% identity), W. koreensis (with 100% identity), and Leuconostoc gelidum or Leuconostoc inhae (with 100% identities), respectively. The same results were obtained for kimchi fermented at −1.5~0 °C for 2~3 months in our previous study [7]. Kim et al. also reported that L. gelidum is a dominant Leuconostoc species in kimchi fermented at 8 °C [26]. Psychrotrophic LAB Lb. sakei and W. koreensis are frequently detected and isolated from kimchi using a routine isolation method for mesophiles because both can grow well Psychrotrophic LAB Lb. sakei and W. koreensis are frequently detected and isolated from kimchi using a routine isolation method for mesophiles because both can grow well at 0 or 30 • C [10,11]. However, L. gelidum or L. inhae can barely grow at 30 • C [14,27]. Although L. gelidum or L. inhae are frequently detected in kimchi using culture-independent methods [6,8], few studies have attempted to characterize these LAB species because they are rarely isolated from kimchi due to the use of 30 • C for isolation experiments [27][28][29].
Kimchi samples, stored at 0~10 • C for 1~3 months, were collected from different locations in South Korea. LAB isolates 1 were identified by determining 16S rRNA gene sequences (1350~1522 bp) and then compared with those in the NCBI Genbank database 2 . In the present study, we isolated and characterized LAB strains corresponding to band c in Figure 1, which possibly were psychrotrophic strains of L. gelidum or L. inhae. As shown in Table 1, 23 LAB strains were isolated from the 16 collected samples and identified based on their 16S rRNA gene sequences-that is, 12 strains of L. gelidum subsp. gasicomitatum, six strains of L. gelidum subsp. aenigmaticum, two strains of each of L. inhae and Lactobacillus algidus, and one strain of Leuconostoc carnosum, which are all psychrotrophic LAB [30,31]. We selected six strains (LAB 11,14,18,19,21,and 22) of L. gelidum subsp. aenigmaticum from the 23 isolates for further investigation, because limited information has been published on L. gelidum subsp. aenigmaticum. The selected six LAB isolates were catalase-negative, Gram-positive, and oval coccishaped cells and formed glossy, ivory-colored colonies. All six LAB isolates assimilated L-arabinose, ribose, D-xylose, D-glucose, D-fructose, D-mannose, α-methyl-Dglucoside, N-acetyl glucosamine, esculine, cellobiose, maltose, sucrose, trehalose, raffinose, β-gentiobiose, D-turanose, 2-keto-gluconate, and 5-keto-gluconate. On the other hand, the six strains differed in terms of carbohydrate assimilation, e.g., LAB 11 was negative for melibiose utilization and LAB 11 and 14 were negative for gluconate utilization (Supplementary Table S2).
According to a recently published reclassification study, L. gelidum comprises three phylogenetically distinct subspecies, L. gelidum subsp. gelidum, L. gelidum subsp. gasicomitatum, and L. gelidum subsp. aenigmaticum [14]. These three subspecies exhibited distinctive phenotypic characteristics with respect to growth in the presence of 6.5% NaCl, hemestimulated aerobic growth, and acid production from amygdalin, arbutin, and salicin. For example, L. gelidum subsp. aenigmaticum did not assimilate amygdalin, arbutin, or salicin, but it assimilated ribose. Furthermore, most of the L. gelidum subsp. aenigmaticum grew in the presence of 6.5% NaCl and the aerobic growth of L. gelidum subsp. aenigmaticum was not stimulated by heme [14]. As shown in Supplementary Table S2 (indicated by shaded region), the six isolated LAB strains showed distinctive characteristics for L. gelidum subsp. aenigmaticum.

Growth at Low Temperature
Growth of the six selected LAB isolates at 0~37 • C in MRS broth was investigated. Typical of the species, L. gelidum grew at 5 • C, but not at 37 • C [14]. As shown in Figure 2, all six LAB grew well at 5 • C and even grew at 0 • C, but not at 37 • C. Maximum cell growth was observed at 15~25 • C, and all six isolates grew better at 5 • C than at 30 • C. Unlike most Leuconostoc, L. gelidum subsp. aenigmaticum, L. gelidum subsp. gasicomitatum, L. gelidum subsp. gelidum, L. inhae, and Lb. algidus are known to have more obligate psychrotrophic characteristics [30]. Kimchi samples, stored at 0~10 °C for 1~3 months, were collected from different locations in South Korea. LAB isolates 1 were identified by determining 16S rRNA gene sequences (1350~1522 bp) and then compared with those in the NCBI Genbank database 2 .
According to a recently published reclassification study, L. gelidum comprises three phylogenetically distinct subspecies, L. gelidum subsp. gelidum, L. gelidum subsp. gasicomitatum, and L. gelidum subsp. aenigmaticum [14]. These three subspecies exhibited distinctive phenotypic characteristics with respect to growth in the presence of 6.5% NaCl, hemestimulated aerobic growth, and acid production from amygdalin, arbutin, and salicin. For example, L. gelidum subsp. aenigmaticum did not assimilate amygdalin, arbutin, or salicin, but it assimilated ribose. Furthermore, most of the L. gelidum subsp. aenigmaticum grew in the presence of 6.5% NaCl and the aerobic growth of L. gelidum subsp. aenigmaticum was not stimulated by heme [14]. As shown in Supplementary Table S2 (indicated by shaded region), the six isolated LAB strains showed distinctive characteristics for L. gelidum subsp. aenigmaticum.
Psychrophilic and psychrotrophic microorganisms can grow at 0 • C. Psychrophiles have optimal growth temperatures below 15 • C and an upper limit of 20 • C, while psychrotrophs (psychrotolerants) grow optimally at 20~25 • C [15,34]. Our results demonstrate that the six strains of L. gelidum subsp. aenigmaticum selected showed typical psychrotrophic growth patterns.
On the other hand, Leuconostoc species including L. citreum and L. mesenteroides (mesophiles), and L. gelidum subsp. gelidum, L. gelidum subsp. gasicomitatum, and L. inhae (psychrotrophs) are readily detected in fermented kimchi [2,6]. Furthermore, Leuconostoc species, mainly L. citreum and L. mesenteroides, isolated from kimchi improve the organoleptic quality and are used as starter cultures to improve sensory qualities and extend shelflife [2]. Meanwhile, psychrotrophic Leuconostoc species, which are detectable in kimchi samples using a culture-independent method [8], have never been used as starter cultures for kimchi fermentation due to their infrequent isolations.
In this study, to investigate the sensory attributes of LAB isolates in fermented food, we performed sensory evaluations of kimchi juices fermented using the LAB isolates. Kimchi juices were prepared and fermented using the six selected LAB isolates as starter cultures. The initial pH value of freshly prepared kimchi juice was pH 5.7. After fermentation, the pH values of LAB juices ranged from 4.01 to 4.23, while the pH value of the control juice was 5.5. Sensory evaluations of fermented kimchi juices (Table 2) showed that the sourness of LAB juices increased according to pH values. Furthermore, the sensory qualities of the juices depended on the starters used. In particular, LAB 21 and LAB 22 juices had fresh tastes and pleasant fermented flavors combined with moderate sourness and a carbonated taste, whereas the other four LAB juices had poorer sensory qualities due to a strong off-flavor (a sewerage-like smell) as compared with control juice. Freshly prepared juice was fermented at 25 • C for 24 h without adding starter (control juice) and after adding each LAB isolate (LAB 11~22 juices). Sensory evaluations were carried out using the prepared samples and rated using a 5-point scale for sourness, fresh taste, carbonate taste, pleasant fermentative smell, and sewerage-like smell (1 = very weak, 3 = moderate, and 5 = very strong) and overall acceptance (1 = very bad, 3 = moderate, and 5 = very good). Means with different letters indicate significant differences (p < 0.05). LAB 21 and LAB 22 showed good sensory properties for kimchi juice fermentation, which suggests that they have potential as useful LAB starter cultures for vegetable or fruit fermentation at low temperatures. Based on the above results, we finally selected LAB 22, which exhibited more obligate psychrotrophic characteristics than LAB 21 (in Figure 2), and designated it L. gelidum subsp. aenigmaticum LS4.

Safety Aspects
Psychrotrophic LAB species have attracted much attention recently [35,36]; however, information about psychrotrophic LAB is lacking. LAB are generally recognized as safe and have a long history of use as starter cultures for various fermented foods [9]. New species and more specific bacterial strains are being sought as novel starter cultures or probiotic candidates; however, the efficacy and safety of new strains must be evaluated.
L. aenigmaticum LS4 did not exhibit αor β-hemolytic activities (data not shown) or any harmful enzyme activities, especially β-glucuronidase and α-chymotrypsin activities (Supplementary Table S3), which may have negative effects in the colon [9]. In API ZYM analysis, L. aenigmaticum LS4 presented β-galactosidase activity (≥40 nmol). However, L. aenigmaticum LS4 (LAB 22) produced a lactose-negative reaction in the API 50 CHL assay (Supplementary Table S2). In vivo, lactose is hydrolyzed to glucose and galactose by β-galactosidase, and thereafter, glucose and galactose are metabolized to produce acids [9]. Thus, we further evaluated the β-galactosidase activity of L. aenigmaticum LS4 and found that its activity was quite low (15.33 Miller units) as compared with other LAB strains (more than 24 Miller units) that produced a lactose-positive reaction in the API 50 CHL assay (data not shown). This result shows that L. aenigmaticum LS4 exerts low β-galactosidase activity; in the API 50 CHL assay, its lactose utilization was too low to register as positive (supplementary Table S2).
Antibiotic resistance testing showed that L. aenigmaticum LS4 was susceptible to all antibiotics tested, except vancomycin (Table 3). According to the technical guidelines issued by EFSA [19], no breakpoint for vancomycin is required for Leuconostoc spp., as the genus Leuconostoc is well known to be intrinsically resistant to vancomycin [9]. In the BA production assay, L. aenigmaticum LS4 did not produce BA(s), histamine, putrescine, or tyramine from histidine, ornithine, or tyrosine (precursor amino acids). In contrast, Lactobacillus sp. ATCC 33222™ and E. faecalis ATCC 29212™, which were used as positive controls for BA production, produced BA(s) from precursor amino acids (Supplementary Figure S1). On the other hand, Leuconostocs are known to produce BA and tyramine from tyrosine [36].
Based on considerations of the above virulence determinants, L. aenigmaticum LS4 does not possess any harmful characteristics such as hemolytic activity, undesirable enzymatic activities, antibiotic resistance, or biogenic amine-producing ability. Thus, it can be reasonably concluded that L. aenigmaticum LS4 can be safely used as a starter culture in food or feed.

Technical Aspects
We investigated the heat, salt (NaCl), and acid/alkali tolerances of L. aenigmaticum LS4 (Table 4). L. aenigmaticum LS4 was stable when exposed to temperatures from −2 to 20 • C (100% survival rate at 48 h). However, no surviving L. aenigmaticum LS4 cells were detected after treatment at 50 to 70 • C for 24 h, and its viability was dramatically reduced after exposure to 30 • C for 48 h and no viable cells were detected after 72 h at 30 • C. 9.40 ± 0.04 azA 9.31 ± 0.01 azB 9.32 ± 0.05 azA 9.10 ± 0.06 ayB 4 • C, pH 4.0, NaCl 3% 9.40 ± 0.04 azA 9.31 ± 0.03 azB 9.33 ± 0.05 azA 9.19 ± 0.04 ayB L. aenigmaticum LS4 cultivated for 24 h in MRS was harvested and resuspended in fresh MRS broth of the same volume (9.40 ± 0.04 log CFU/mL). Thereafter, L. aenigmaticum LS4 tolerances of pH, temperature, or NaCl were determined as described in Materials and Methods. Means with different letters are significantly different (p < 0.05) in the same column (a~c), in the same row (w~z), and in the same column, indicated by shading (A~D). 1 N.D: Not detected.
L. aenigmaticum LS4 was unaffected by 0~5% NaCl for 24 h at 25 • C, but its viability was dramatically reduced at NaCl concentrations > 7%; no viable cells were detected after treatment with 9% NaCl for 72 h. Interestingly, cell viabilities after treatment with 3 to 5% NaCl were significantly greater than after treatment with 0 to 1% NaCl for 72 h. The optimum NaCl concentration range with respect to cell viability after exposure for 72 h was 3 to 5% NaCl (Table 4).
L. aenigmaticum LS4 was reasonably stable in the pH range 6.0 to 8.0 for 24 h at 25 • C, but no surviving cells were observed after treatment with pH 2.0 or 10.0 for 24 h (Table 4). Cell viability steadily decreased after 24 h of exposure to pH 4.0, and at 72 h, only 1 to 0.1% of cells remained viable as compared with cells exposed to pH values of 6.0 to 8.0. When we compared the results of other LAB strains, namely the L. citreum, L. mesenteroides, Lb. sakei, and Lb. plantarum strains [7,9,38,39], L. aenigmaticum LS4 showed similar or slightly weaker acid tolerances than L. citreum and L. mesenteroides (heterofermentative LAB) but significantly poorer acid tolerances than Lb. sakei and Lb. plantarum (homofermentative LAB) strains. On the other hand, Jung et al. reported that Leuconostoc spp., such as L. gelidum subsp. gelidum and L. gelidum subsp. gasicomitatum, are more acid-tolerant than L. citreum and L. mesenteroides, because L. mesenteroides and L. citreum are dominant during early fermentation, while L. gelidum, L. gasicomitatum, and Lb. sakei are dominant at the late kimchi fermentation stage, as determined by culture-independent analysis [8]. However, they did not determine the acid tolerances of these LAB. The results of our study indicate that the acid tolerance of L. aenigmaticum LS4 is not as high as that of the homofermentative LAB Lb. plantarum [7].
We further investigated the effect of low temperature (4 • C) on the cell viability of L. aenigmaticum LS4 at low pH (pH 4.0). Cell viability after exposure to 4 • C/pH 4.0 in MRS was significantly greater than that observed at 25 • C/pH 4.0 in MRS and almost the same as (or slightly lower than) that observed at 4 • C/pH 6.5 in MRS. Notably, the cell viability of L. aenigmaticum LS4 at pH 4.0 was significantly improved by incubation at 4 • C (3.03 log CFU/mL at pH 4.0/25 • C/72 h vs. 9.10 log CFU/mL at pH 4.0/4 • C for 72 h vs. 9.38 log CFU/mL at pH 6.5/4 • C/72 h) (shaded regions in Table 4). These results show that L. gelidum and L. gasicomitatum dominate during the late stage of kimchi fermentation [8,13] because kimchi is stored at temperatures below 10 • C for 1~4 months under acidic conditions (pH~4.0). In other words, this occurred because psychrotrophic L. gelidum or L. gasicomitatum have lower optimum growth temperatures than the mesophiles L. citreum or L. mesenteroides [30,40] and better tolerate low temperatures [41]. Stress tolerance results showed that L. aenigmaticum LS4 best tolerated temperatures of −2 to 10 • C and that it achieved a 100% survival rate.
LAB encounter various environmental stresses (e.g., thermal, acid/alkali, salt) during food fermentation. LAB intended for use as starter cultures in the food industry must be able to survive and retain their bioabilities under industrial conditions and in fermented food products. The results in Table 4 demonstrate that L. aenigmaticum LS4 is a suitable starter culture for vegetable or fruit fermented at temperatures below 10 • C that can maintain thermally instable nutrients or flavors.

Functional Aspects
EPS production on sucrose medium by the six LAB isolates (LAB 11,14,18,19,21,and 22) was observed after culture for 48 h at 25 • C ( Figure 3) and was greater for LAB 18,19,21, and 22 than for LAB 11 and 14. LAB EPSs are important bioproducts that have been shown to improve shelf-life, to enhance techno-functional applications, and to provide health benefits [42]. EPSs can be classified as heteropolysaccharides (HePSs) and monopolysaccharides (HoPSs) based on their monosaccharide compositions and biosynthetic mechanisms [42]. We determined the monosaccharide contents and the molecular weights of the EPS produced by L. aenigmaticum LS4 (LAB 22) by HPLC and GPC, respectively. L. aenigmaticum LS4 produced EPS composed of glucose only (Figure 3), with molecular weight ≈ 3.7 × 10 6 Da (data not shown), which is glucan. The molecular weights of the glucans produced by other LAB strains, including Leuconostoc, Lactobacillus, Pediococcus, and Weissella, ranged from 10 3 to 10 7 Da [43]. EPS is not used as an energy source by producer microorganisms; rather, it protects microbial cells exposed to harsh conditions (e.g., acidic and bilious conditions) [9,40]. We believe that EPS production could improve the techno-functional applications of L. aenigmaticum LS4 in the food industry. A considerable number of investigations on the EPSs produced by Leuconostoc have been conducted due to their wide-ranging applications in the food, medical, and industrial fields. However, most are related to mesophilic Leuconostoc strains such as L.mesenteroides, L. citreum, or L. pseudomesenteroides [44].
(e.g., acidic and bilious conditions) [9,40]. We believe that EPS production could improve the techno-functional applications of L. aenigmaticum LS4 in the food industry. A considerable number of investigations on the EPSs produced by Leuconostoc have been conducted due to their wide-ranging applications in the food, medical, and industrial fields. However, most are related to mesophilic Leuconostoc strains such as L.mesenteroides, L. citreum, or L. pseudomesenteroides [44]. As shown in Table 5, L. aenigmaticum LS4 showed strong antibacterial activities against food-borne pathogenic Bacillus cereus, Escherichia coli, Listeria monocytogenes, Pseudomonas aeruginosa, Salmonella enterica serovar. Typhi, and Vibrio parahaemolyticus. L. aenigmaticum LS4 also showed antifungal activities against food spoilage fungi, including Aspergillus flavus, Aspergillus fumigatus, Aspergillus nidulans, Aspergillus ochraceus, and Penicillium roqueforti. L. aenigmaticum LS4 did not exhibit antimicrobial activities against bacteria Micrococcus luteus and Staphylococcus aureus. Its antibacterial activities were stronger than its antifungal activities.   To determine whether its antimicrobial activities were due to the organic acids produced, organic acids produced by L. aenigmaticum LS4 in the MRS culture were quantified by IC analysis: 8642.54 mg/L lactic acid, 43.73 mg/L phenyllactic acid, and 14.66 mg/L fumaric acid. Thereafter, the antimicrobial activities of the culture filtrate of L. aenigmaticum LS4 and the organic acid mixture in L. aenigmaticum LS4 culture quantified were simultaneously assayed. As shown in Table 5, the antibacterial activities of the culture filtrate of L. aenigmaticum LS4 were significantly higher than those of the organic acid mixture quantified. However, the antifungal activities of the L. aenigmaticum LS4 culture filtrate and the organic acid mixture were almost the same. These results indicated that the antifungal activities of L. aenigmaticum LS4 probably originate from the organic acids produced, whereas its antibacterial activities originate from some other antimicrobial(s) (e.g., bacteriocin) and the organic acids produced. We then examined the antimicrobial activities of the culture filtrate treated with protease ( Table 5). As was expected, protease treatment did not affect the antifungal activities of the culture filtrate. However, the antibacterial activities of the culture filtrate were significantly reduced by protease treatment and were similar to those of the organic acid mixture. These results confirmed the proteinaceous nature of the bacteriocin produced by L. aenigmaticum LS4 and imply its digestion in the gastrointestinal tracts of humans and animals.
EPS production and the antimicrobial activities of L. aenigmaticum LS4 as a starter culture would contribute to the health benefits and safety of fermented foods.
Cold shock protein (CSP), DEAD-box RNA helicase, and ribonuclease (RNase) are commonly known cold-shock response gene families in L. gasicomitatum, Lactococcus piscium, and Paucilactobacillus oligofermentans (formerly Lactobacillus oligofermentans), which are all psychrotrophic LAB [45]. In addition, ribosomal protein, tRNA and rRNA modification, and ABC and efflux MFS transporter genes have been suggested to be components of the cold-shock response machinery [45]. In this analysis of genes related to the coldshock response by psychrotrophic L. aenigmaticum LS4, we detected CSP, DEAD-box RNA helicase, and RNase genes as well as genes for efflux ABC transporter, ribosomal protein, and rRNA/rRNA modification (Supplementary Table S4), which have been reported to be related to adaptation or the active growth required for bacterial survival at low temperatures [45,46]. Temperature is one of the major stresses that all living microorganisms must face because food-related bacteria including LAB are repeatedly exposed to low temperatures. The cold-shock-response-related proteins in Supplementary Table S4 have been identified in a variety of microorganisms (not only in psychrotrophs but also in Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Institutional Review Board of Chosun University (IRB # 2-1041055-AB-N-01-2020-29).
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.