2022 Volume 28 Issue 3 Pages 267-273
Natto is a fermented food made from soybeans using Bacillus subtilis natto. B. subtilis is present in honey. In this study, we evaluated the genetic and enzymatic characteristics of B. subtilis obtained from commercial honey and its suitability for natto fermentation. Three bacterial strains isolated from honey samples of different origins were found to be highly homologous to B. subtilis via 16S rDNA analysis. Similar to B. subtilis natto, all three isolates were biotin auxotrophs and possessed two insertion sequences. Genotyping by random amplified polymorphic DNA-polymerase chain reaction and enzyme activity analysis showed that the isolates were genetically different from the commonly used natto strains Miyagino, Takahashi, and Naruse. The three isolates are highly suitable for natto fermentation. These results indicate that honey is a source of B. subtilis natto that is potentially useful in the food industry.
Natto is one of the most widely consumed fermented foods in Japan. It is made from soybeans using Bacillus subtilis natto as a starter strain. Natto is rich in viscosity, composed of γ-polyglutamic acid (γ-PGA) and levan, and has a unique aroma derived mainly from pyrazines and short-chain fatty acids (Hara et al., 1984; Tanaka et al., 1998). The characteristics of different natto products are greatly influenced by the variety of starter strain used, in addition to fermentation conditions and the variety of the soybeans used. B. subtilis natto is classified as B. subtilis subsp. subtilis (Nagai, 2014) and efficiently produces extracellular protease, sucrase, and cellulase, which are important factors affecting the physical properties and taste of natto products. B. subtilis natto is also characterized by biotin auxotrophism, phage infectivity, and the possession of insertion sequences (IS) (Kubo et al., 2011).
Commercial production of natto is implemented using one of the three major B. subtilis natto starter strains: Takahashi, Naruse, and Miyagino, except for some products. To maintain the diversity of natto products, discovering novel strains of B. subtilis natto is important. To date, B. subtilis natto has been isolated primarily from rice straw. It has been recently reported that soybean-fermented foods in Southeast Asia, such as Thua nao, potentially harbor B. subtilis natto (Inatsu et al., 2006). However, there are few reports on its isolation from other foods or environmental sources.
Honey severely limits the presence of microorganisms owing to its acidity, high osmotic pressure caused by its high sugar concentration, and the presence of hydrogen peroxide. B. subtilis forms resistant endospores in response to harsh environmental conditions. Thus, B. subtilis can survive in the honey environment. In fact, B. subtilis has been found in honey, its origin is thought to be from pollen and/or honeybees (Gilliam, 1997).
Extensive research has been conducted on B. subtilis associated with honey, including its effectiveness as a probiotic and its use in the production of antimicrobial agents (Hamdy et al., 2017; Amin et al., 2020); however, its suitability for natto fermentation has not been reported. In this study, we isolated and identified B. subtilis from honey and evaluated its suitability for natto fermentation.
Bacterial strains and culture conditions Brain-heart infusion medium (BHI; Becton Dickinson and Company, San Jose, CA) was used to isolate the bacteria from honey and maintain the bacterial culture. Isolated strains, soybean fermentation starter B. subtilis natto (Miyagino, Naruse, and Takahashi strains), and B. subtilis wild type (Marburg 168; Burkholder and Giles, 1947; Spizizen, 1958) were cultured at 37 °C for 20 h under aerobic conditions. Biotin auxotrophic tests were performed on E9 agar medium without L-glutamic acid (Birrer et al., 1994) supplemented with biotin (0.5 µg/ml) and 1.5% agar (w/v).
Isolation of B. subtilis from honey samples Three types of honey were used as isolation sources: one honey sample was from Japan, and the other two samples were mixtures from China and Argentina. Honey (100 µg) was spread on BHI agar plates containing 1.5% agar (w/v). The plates were incubated overnight under aerobic conditions at 37 °C. Wrinkled and raised colonies (characteristic of B. subtilis natto) were isolated. The isolates were purified on BHI agar plates and used for subsequent experiments.
Biotin auxotrophy The natto starter strains and isolated strains were cultured in BHI liquid medium. After centrifugation of 1 ml of the culture medium, the cell pellet was suspended in 1 mL of sterile water. Briefly, 5 µl of the cell suspension was spotted on E9 medium with 1.5% agar (w/v). After three days of aerobic incubation at 37 °C, growth on the agar medium was visually confirmed.
DNA extraction DNA extraction from bacterial cultures was performed using the DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany).
16S rDNA sequencing Strains isolated from honey were identified using 16S rDNA sequences. A polymerase chain reaction (PCR) fragment of approximately 1.5 kb was amplified using universal primers (27F and 1512R) and purified using the QIAquick PCR purification kit (Qiagen). The DNA sequences were confirmed in duplicate by sequencing both strands. Partial 16S rDNA sequences (approximately 860 bp) were analyzed by alignment with the GenBank database using BLAST.
Amplification of bioF and IS genes The primers used to amplify the bioF and IS genes are listed in Table 1. The genomic information of B. subtilis BEST195 (http://natto-genome.org/) was used to design the primers required for gene amplification. The PCR thermal profile for the amplification of the genes was 98 °C for 1 min, 56 °C for 1 min, and 70 °C for 1 min for 30 cycles. The PCR products were electrophoresed on 1.5% agarose and evaluated by the presence of bands after staining with ethidium bromide.
| Primer name | Sequnce (5′→3′) | Amplicon | Reference | |
|---|---|---|---|---|
| 27F | AGAGTTTGATCCTGGCTCAG | 165 rDNA | Lane, 1991 | |
| 1512R | ACGGCTACCTTGTTACGACT | Lane, 1991 | ||
| BEST195 bioF 200F | AGATTGCCAGCTTTAAGCGGAC | bioF | In this study | |
| BEST195 bioF 963R | CGTGATTGTAATTCGAATACGGC | In this study | ||
| IS4Bsu1 290F | CTCAATGAAACCGAGAGTCTGC | IS4Bsu1 | In this study | |
| IS4Bsu1 1097R | TATGTACGCCTTGTTCGCTC | In this study | ||
| IS4Bsu256 51F | TAACAAGATTATCGAACAGTATCAGCC | IS4Bsu256 | Kimura and Itoh., 2007 | |
| IS4Bsu256 1063R | AGCTAGAGTGGATACTTTCAACGGC | Kimura and Itoh., 2007 | ||
| RAPD forward | CTC AAT GAA ACG GAG ATC TGC | RAPD (IS4Bsu1) | In this study | |
| RAPD reverse | ||||
| OPA-9 | GGGTAACGCC | Hamid et al (2015) | ||
| OPAH-17 | CAGTGGGGAG | Hamid et al (2015) | ||
| OPBB-4 | ACCAGGTCAC | Hamid et al (2015) | ||
| OPH-13 | GACGCCACAC | Hamid et al (2015) | ||
| OPH-18 | GAATCGGCCA | Hamid et al (2015) | ||
| RI-3 | GTCCGTGAAC | Hamid et al (2015) | ||
| RI-16 | GTCGCCGTCA | Hamid et al (2015) | ||
| RAPD2 | CCGAGTCCA | Hosoi (2010) | ||
| RAPD3 | GTTTCGCTCC | Hosoi (2010) | ||
| RAPD4 | AAGAGCCCGT | RAPD (random) | Hosoi (2010) | |
| RAPD5 | CCGCAGCCAA | Hosoi (2010) | ||
| RAPD6 | TGCCGAGCTG | Hosoi (2010) | ||
| RAPD7 | AATCGGGCTG | Hosoi (2010) | ||
| RAPD8 | CAATCGCCGT | Hosoi (2010) | ||
| RAPD9 | GGTGATCAGG | Hosoi (2010) | ||
| RAPD10 | CCGGCGGCG | Hosoi (2010) | ||
| RAPD11 | AGTCGGGTGG | Hosoi (2010) | ||
| RAPD12 | AGGGGGTTCC | Hosoi (2010) | ||
| OPB11 | CTAGACCCGT | Pedroso (2012) | ||
Randomly amplified polymorphic DNA (RAPD) PCR analysis PCR amplification was performed using one forward primer and a primer set of 19 different reverse primers (Table 1). Briefly, 10 µl of PCR solution was prepared as follows: 3 µL Premix Taq™ (TaKaRa Taq™ Version 2.0) (TAKARA, Tokyo), 5 ng DNA, 5 µM primer, and scaled-up to 10 µL with PCR-grade water. The amplification conditions were as follows: an initial denaturation step at 94 °C for 2 min, 40 cycles of 92 °C for 1 min, 39 °C for 1 min, and 72 °C 2 min, and a final elongation step at 72 °C for 5 min. The PCR products were confirmed using the capillary electrophoresis system MCE-202 MultiNA (Shimazu, Kyoto, Japan). To obtain reproducible, accurate, and distinct banding patterns, all amplifications were independently repeated at least three times. Genotypes were scored as “1” if the band was present and “0” if it was absent. Phylogenetic trees were constructed using the statistical analysis software “R” (version 3.4.3) based on the unweighted pair group method with arithmetic mean (UPGMA method).
Preparation of crude enzyme solutions from bacterial cultures Bacterial cultures (150 µL) were inoculated into 15 mL of fresh BHI medium supplemented with 1% sucrose (w/v). After aerobic shaking at 37 °C for 20 h, the culture (15 mL) was centrifuged and precipitated with 80% (w/v) ammonium sulfate at 4 °C for more than 3 h. The precipitated fraction was dissolved in 15 mL of 100 mM phosphate buffer (pH 7.4), subjected to ultrafiltration (10 kDa; Merck, Germany), and the resulting fraction (> 10 kDa) was suspended in 1.5 ml of 100 mM phosphate buffer (pH 7.4).
Crude protein and enzyme activity The protein content of the crude enzyme solutions was determined using a protein assay kit (Thermo Fisher Scientific; Waltham, MA, USA) with bovine serum albumin as the standard. Proteolytic activity was measured via azo-casein hydrolysis (Iwamoto et al., 2018). One unit of proteolytic activity was defined as the amount of enzyme required for an increase in the absorbance at 440 nm (ΔA440/min/mL).
Levansucrase activity was evaluated by the amount of glucose released from sucrose. Briefly, 50 µl of 1% sucrose solution (w/v) and 50 µL of crude enzyme solution was mixed and incubated at 37 °C for 30 min. The amount of released glucose was measured using a Saccharose/D-Glucose/D-Fructose F-kit (JK International, Tokyo, Japan). One unit of sucrase activity was defined as the amount of enzyme required to release 1 µg of glucose (Δµg glucose/min/mL).
The activity of endoglucanase was evaluated based on the amount of reducing sugars released from carboxymethyl cellulose (CMC) as a substrate. Briefly, 200 µL of 1% CMC solution dissolved in 50 mM acetate buffer (pH 6.0) was mixed with 50 µL of crude enzyme solution and incubated at 50 °C for 15 min. The amount of reducing sugars released was measured via the 3,5-dinitrosalicylic acid method. One unit of endoglucanase activity was defined as the amount of enzyme that releases 0.1 mmol of glucose per minute (Δ100 nmol glucose/min/mL).
Natto fermentation Soybean fermentation was done at 37 °C for 44 h under approximately 90% relative humidity as previously reported (Yokoyama et al., 2021). Samples were stored at −30 °C until further use.
γ-PGA extraction and quantification The γ-PGA amount in fermented soybean samples was determined via the cetyltrimethylammonium bromide method (Kanno and Takamatsu, 1995) and calculated as the glutamic acid equivalent (Kiuchi, 2010).
Statistical analysis The means of at least three independent experiments were calculated along with their standard deviations. The data were analyzed using the Tukey-Kramer method.
B. subtilis-like colonies were isolated from three types of commercial honey. The three isolates from one of Chinese and Argentinian mixture honey, other of Chinese and Argentinian mixture honey, and Japanese honey samples were designated H1-1, H3-1, and H11-1, respectively, and all exhibited high homology to B. subtilis via 16S rDNA analysis of partial sequences (Table 2). In B. subtilis, biotin synthesis is regulated by a bio operon (bioWAFDBI) (Sasaki et al., 2004). B. subtilis natto cannot synthesize biotin because it lacks approximately 50 bp of the bioF gene encoding 8-amino-7-oxononanoate synthase. This characteristic is a critical difference between B. subtilis natto and B. subtilis wild type. All three isolated strains required biotin (Table 2). Each bioF fragment amplified from the three isolated strains was shorter than expected for the B. subtilis wild type and was consistent with the natto starter strains Miyagino, Takahashi, and Naruse (Table 2). The biotin requirement of the Naruse strain, an existing natto starter, could not be evaluated because it showed a false-negative reaction in E9 medium. The existing natto starters had insertion sequences IS4Bsu1 and IS4Bsu256; however, the B. subtilis wild type lacked these sequences. The isolates H1-1, H3-1, and H11-1 had two different insertion sequences (Table 2), indicating their similarity with the existing natto starter strains.
| Strain | 16S rDNA (Identify) | Biotin Auxotrophy | bioF (Amplification length) | Insertion Sequence (Amplification length) | |
|---|---|---|---|---|---|
| IS4Bsu1 | IS4Bsu256 | ||||
| B. subtilis | |||||
| wild type (Marburg 168) | N.T. | N.T. | +(810 bp) | N.D. | N.D. |
| natto Miyagino | N.T. | Positive | +(760 bp) | + (800 bp) | + (1,000 bp) |
| natto Naruse | N.T. | Unrecognizable | +(760 bp) | + (800 bp) | + (1,000 bp) |
| natto Takahashi | N.T. | Positive | +(760 bp) | + (800 bp) | + (1,000 bp) |
| Honey isolates | |||||
| H1-1 | B. subtilis (859 bp/860 bp) | Positive | +(760 bp) | + (800 bp) | + (1,000 bp) |
| H3-1 | B. subtilis (860 bp/860 bp) | Positive | +(760 bp) | + (800 bp) | + (1,000 bp) |
| H11-1 | B. subtilis (860 bp/861 bp) | Positive | +(760 bp) | + (800 bp) | + (1,000 bp) |
N. T., not tested.
N. D., not detected.
The “+” indicates the presence of a band in the agarose gel.
The RAPD-PCR method is widely used to study genotypic similarities and variations (Sarkar et al., 2002; Inatsu et al., 2006; Hosoi, 2010). We used the RAPD-PCR method to evaluate the genotypic similarity between the isolated strains and the existing natto starters. In B. subtilis natto, the number of copies of IS4Bsu1 in genomic DNA varies among strains (Nagai et al., 2000). RAPD-PCR analysis was performed using 19 primer sets, which included a forward primer based on the IS4Bsu1 gene sequence and 19 different random reverse primers. All 19 primer sets (Table 1) exhibited good amplification, and 185 reproducible and clear bands were obtained in a size range of 40–2,000 bp (data not shown). Polymorphisms were observed in 48% of 185 bands. The polymorphism rate of the primer OPBB4 was 75%, the highest among the primers used in this study. Eight primer sets (RAPD2, RAPD4, RAPD6, RAPD11, OPA9, OPB11, OPAH17, and OPH18) had low polymorphism rates (<50%) and were excluded from the dendrogram analysis. In the dendrogram generated from the genetic distance using 103 bands, the genotypes were classified into three groups. The three starter strains were included in the first group, the H11-1 strain in the second group, and H1-1 and H3-1 strains in the third group (Fig. 1). The genetic distances of the six strains ranged from 0.15 to 0.84, with an average of 0.65. The highest genetic distance was observed between Miyagino and Naruse. The lowest genetic distance was observed between H11-1 and H3-1. These results suggest that the three strains isolated from honey are genetically different from the known natto strains used in commercial natto production.
Unweighted pair group method with arithmetic mean dendrogram illustrating genetic distance of six strains based on randomly amplified polymorphic DNA-polymerase chain reaction method.
The enzymatic activity of B. subtilis derived from fermented soybean food varies among strains (Chantawannakul et al., 2002; Inatsu et al., 2006). Therefore, to investigate the characteristics of the isolated strains, extracellular protease activity, levansucrase activity, and endoglucanase activity, which affect the physical properties and taste of natto products, were evaluated using the crude enzyme solutions derived from culture supernatants. The results are shown in Table 3. The protease activity of strain H11-1 was higher than that of strains H1-1 and H3-1 and was comparable to that of B. subtilis natto Takahashi. The levansucrase activity was not significantly different among the isolates and was comparable to that of the existing starter bacteria. The characteristics of endoglucanase productivity of the three strains isolated from honey were similar to those of the proteases, with high productivity of isolate H11-1 and comparable productivities of isolates H1-1 and H3-1.
| Strain | Protease (unit/mg protein) | Levansucrase (unit/mg protein) | Endoglucanase (unit/mg protein) |
|---|---|---|---|
| B. subtilis | |||
| Wild type (Marburg 168) | 1.02±0.31b | 0.36±0.08b | 0.07±0.05d |
| natto Miyagino | 1.38±0.2b | 3.38±0.56a | 2.82±0.38c |
| natto Naruse | 1.23±0.58b | 3.26±1.56a | 0.20±0.12d |
| natto Takahashi | 3.95±2.02a | 5.05±1.04a | 6.56±0.08a |
| Honey isolates | |||
| H1-1 | 1.50±0.61b | 3.98±1.86a | 3.15±0.71c |
| H3-1 | 1.37±0.46b | 3.69±1.20a | 3.16±0.88c |
| H11-1 | 2.75±1.52b | 3.86±1.24a | 4.33±0.85b |
Different letters indicate significant statistical differences by Tukey–Kramer method (p < 0.05).
The three isolated strains were evaluated for their suitability for natto fermentation. B. subtilis natto Miyagino, Takahashi, and Naruse were used as controls. The natto produced by the new isolates exhibited a uniform biofilm on the surface of soybeans and the characteristic smell of natto. These characteristics were similar to those of the starter bacteria (data not shown). Both types of natto, the ones made using the new isolates and those made using the commonly used starter bacteria, exhibited sufficient stringiness and viscosity. The amount of γ-PGA in the natto products was classified into three categories (Fig. 2). Natto produced by H3-1 and H11-1 strains contained the same amount of γ-PGA as that produced by the Miyagino strain, and natto produced by the H1-1 strain contained considerably more γ-PGA than that produced by the H3-1 and H11-1 strains. These results showed that the isolates were highly suitable for fermentation.
γ-PGA quantity in the natto product. “Strain name” on the x-axis indicate the natto product made with each strain. Data represent average values from three independent experiments. Different letters indicate significant statistical differences by Tukey–Kramer method (p < 0.05).
The amount of γ-PGA in natto products is closely associated with the protease productivity of the starter bacteria (Kada et al., 2013; Yokoyama et al., 2021). In the present study, there was no correlation between protease productivity in the culture medium and the amount of γ-PGA during natto production (Fig. 2, Table 3). Kimura et al. (2004) reported that the amount of γ-PGA is strongly influenced by synthetic and degradative enzymes. In the future, a comprehensive evaluation of the synthesis and degradation of γ-PGA is necessary.
RAPD-PCR analysis of genomic DNA revealed that the honey isolates were genotypically different from the existing natto starters. Thus, honey isolates may have different amounts of metabolites compared to the existing species, providing natto made using these isolates unique taste, texture, and other physical properties. We will need to evaluate the chemical compositions and sensory evaluation of the natto made from the B. subtilis natto obtained in this study.
In this study, three strains of B. subtilis natto were isolated from three different types of honey. In addition to this, we attempted to isolate B. subtilis natto from ten other types of honey, but we were unable to isolate B. subtilis natto (data not shown). In general, the types of cells and the number of colonies obtained vary depending on the culture conditions. In the future, more B. subtilis natto isolates could be obtained by examining various culture conditions.
In general, bacteria are unable to grow when their water activity is below 0.6. The water activity of honey in the present study was around Aw = 0.55 (data not shown). Therefore, bacteria in the honey were considered to be in the form of spores.
Natto and other soybean-fermented foods (e.g., thua nao and kinema) are primarily consumed in Southeast Asia and East Asia. Future studies on honey from more diverse regions are expected to lead to a better understanding of the diversity of B. subtilis natto and the food culture of fermented soybeans.
Acknowledgements B. subtilis Marburg 168 was provided by the National BioResource Project (NIG, Japan). This work was partly supported by a Nihon University College of Bioresource Science Research Grant for 2020.
Conflict of interest There are no conflicts of interest to declare.
