Complete Nucleotide Sequence Analysis of a Novel Bacillus subtilis-Infecting Bacteriophage BSP10 and Its Effect on Poly-Gamma-Glutamic Acid Degradation

While the harmful effects of lactic acid bacterial bacteriophages in the dairy industry are well-established, the importance of Bacillus subtilis-infecting bacteriophages on soybean fermentation is poorly-studied. In this study, we isolated a B. subtilis-infecting bacteriophage BSP10 from Meju (a brick of dried fermented soybean) and further characterized it. This Myoviridae family bacteriophage exhibited a narrow host range against B. subtilis strains (17/52, 32.7%). The genome of bacteriophage BSP10 is 153,767 bp long with 236 open reading frames and 5 tRNAs. Comparative genomics (using dot plot, progressiveMauve alignment, heat-plot, and BLASTN) and phylogenetic analysis strongly suggest its incorporation as a new species in the Nit1virus genus. Furthermore, bacteriophage BSP10 was efficient in the growth inhibition of B. subtilis ATCC 15245 in liquid culture and in Cheonggukjang (a soybean fermented food) fermentation. Artificial contamination of as low as 102 PFU/g of bacteriophage BSP10 during Cheonggukjang fermentation significantly reduced bacterial numbers by up to 112 fold in comparison to the control (no bacteriophage). Moreover, for the first time, we experimentally proved that B. subtilis-infecting bacteriophage greatly enhanced poly-γ-glutamic acid degradation during soybean fermentation, which is likely to negatively affect the functionalities of Cheonggukjang.


Introduction
Bacillus subtilis is a Gram-positive, endospore-forming bacterium commonly found in soil and the human gut [1]. It is an industrially important microorganism used for its probiotic activities in humans [2] and for commercial enzyme production [3]. In addition, B. subtilis is one of the major fermenting bacteria in soybean-based fermented foods such as Korean Cheonggukjang (a fast-fermented soybean product), Deonjang (fermented soybean paste), Gochujang (hot red pepper paste), and Japanese Natto [4,5]. These types of soybean-based fermented foods are an integral part of the diet of Asian peoples and the market values of these products are increasing day by day [4].
A bacteriophage is a virus that infects host bacteria. Due to recent antibiotic resistance problems, much attention has been paid to the control of harmful bacteria using a bacteriophage [6]. However, the bacteriophage infecting beneficial bacteria are as important as those killing pathogenic bacteria. It is well established that for lactic acid bacteria bacteriophage infection could delay or abort the dairy fermentation process and affect the quality, flavor, and texture of the final products [7,8]. In order to

High-Titer Bacteriophage Preparation
Plate elution, large-scale preparation in liquid culture, PEG (polyethylene glycol) precipitation, and CsCl gradient ultracentrifugation of bacteriophage BSP10 were performed as described previously [6], and resulted in a high titer bacteriophage stock preparation (10 11−12 PFU/mL).

Host Range Analysis by Dotting and Plating Method
Host range analysis was performed initially by dotting 5 µL of PEG-precipitated bacteriophage sample (diluted in SM buffer at a final concentration of 10 7 PFU/mL) on the TA soft agar containing indicator bacteria. Host infectivity of bacteriophage BSP10 was further confirmed by a plaque forming assay. Briefly, 100 µL of bacteriophage BSP10 (decimally diluted in SM buffer) was mixed with TA soft agar containing 300 µL of the bacteria (dotting positive) and poured into pre-solidified TSA plates followed by incubation at 37 • C. After 24 h incubation, the plaque formations were monitored. The EOP (efficiency of plating, %) was calculated using the following formula, ([average PFU on target bacteria/average PFU on host bacteria B. subtilis ATCC 15245] × 100). EOP was classified into three categories, high (>50%), medium (10-50%), and low (<10%).

Transmission Electron Microscopy
Dialyzed CsCl-purified bacteriophage stock was used for transmission electron microscopy, as described earlier with some modifications [15]. Briefly, 10 µL of dialyzed CsCl-purified bacteriophage BSP10 was spotted on the carbon-coated copper grid and incubated for 5 min. The excess solution was removed with filter paper, and the grids were negatively stained with 10 µL of 2% uranyl acetate (Sigma, St. Louis, MO, USA). The samples were visualized by Bio-TEM (Hitachi, Tokyo, Japan) with an acceleration voltage of 100 kV and 200,000× magnification.

One Step Growth Assay
A one-step growth assay was performed as previously described with some modifications [16]. To summarize, the culture of B. subtilis ATCC 15245 was grown overnight, added to 25 mL of fresh TA media (1% inoculum) and incubated at 37 • C. When the optical density at 600 nm (OD 600 ) was reached at 0.3 (approximately 1 × 10 7 CFU/mL), the bacteriophage was added at MOI of 0.01 and incubated for 5 min, followed by centrifugation at 26,200× g for 10 min. The pellet was resuspended in the same volume of fresh TA medium. Then 1 mL of this mixture was aliquoted into micro-centrifuge tubes (1.5 mL) followed by incubation at 37 • C with shaking (160 rpm). At every 10 min, one tube was taken out from the incubator and centrifuged at 12,300× g for 3 min. The supernatant was then collected and used for PFU counting by using TA soft agar.

Genomic DNA Isolation and Sequencing
Genomic DNA was extracted using PEG precipitated bacteriophage stock (filter-sterilized) and an ExgeneTM Cell SV DNA purification kit (GeneAll Biotechnology Co., LTD., Songpa-gu, Korea) after DNase and SDS treatment. Whole genome sequencing was done by Macrogen Inc., Geumcheon-gu, Korea using a MiSeq sequencing system (Illumina, San Diego, CA, USA). Raw reads were trimmed and de novo assembled using A5-miseq pipeline with approximately 575× depth of coverage among the 297,028 reads.
The whole genome sequence of Bacillus bacteriophage BSP10 was deposited in GenBank (Accession number MF422185).
The whole genome sequence of bacteriophage BSP10 was analyzed using the CLuster Analysis of Sequences (CLANS) software package [29] and Gegenees 2.0.0 (tBLASTx method, fragment size −50, step size −25) [30]. The dendrogram was generated by SplitsTree 4.13.1 [31], using the neighbor-joining method based on the similarity matrix generated by Gegenees. In addition, the phylogenetic position of bacteriophage BSP10 was determined by using a major capsid protein, a terminase large subunit, and metallophosphatases [32].

Inhibition of Bacterial Growth in Liquid Culture
Liquid culture inhibition of bacteriophage BSP10 was performed as described by Shin, Bandara, Shin, Ryu, and Kim [6], with some modifications. Briefly, 1% of overnight grown B. subtilis ATCC 15245 was added to 25 mL fresh TSB media and incubated at 37 • C with shaking (160 rpm). Bacteriophage BSP10 was added at a multiplicity of infection (MOI) of 1 and 0.1 when OD 600 was approximately 0.3. Bacterial growth (either colony forming units [CFU] counts or OD 600 using spectrophotometer [Biochrom Libra S22 Visible Spectrophotometer, Cambourne, Cambridge, UK]) and plaque forming units (PFU) were monitored at the designated time points for 72 h. The experiment was performed for three times and the values were represented as the mean ± Standard Deviation (SD).

Inhibition of Bacterial Growth in Cheonggukjang
B. subtilis ATCC 15245 growth inhibition by bacteriophage BSP10 was evaluated by modifying the method used for B. cereus inhibition by bacteriophage in Cheonggukjang [6]. Washed soybeans were immersed in distilled water for 10 h at 4 • C. After removing the water, the beans were air-dried at room temperature for 30 min. Ten grams of beans was put into a 100 mL conical flask and autoclaved at 121 • C for 15 min. After cooling the flask at room temperature, B. subtilis ATCC 15245 was added at a concentration of 10 4 CFU/g (which is the standard concentration used in factories) in 1 mL of TA, followed by artificial contamination of bacteriophage BSP10 at 10 2 , 10 4 , and 10 6 PFU/g. The flasks were incubated at 37 • C for 72 h without any perturbation. B. subtilis growth was monitored using TSA and the bacteriophage count was checked using TA soft agar, as stated above, every 24 h. Data were statistically analyzed by using one-way ANOVA (Analysis of variance) and Duncan's multiple range test in Sigma Plot 12.5 (Systat Software, San Jose, CA, USA) to determine the significant relationship between the mean values. The experiment was carried out three times and the values were represented as the mean ± SD.
2.12. Poly-Gamma-Glutamic Acid (γ-PGA) Analysis in Cheonggukjang γ-PGA was extracted by modifying the method described by Zeng, et al. [33]. Briefly, 10 g of Cheonggukjang (24 h incubation sample) was mixed with 20 mL of deionized water followed by shaking at 160 RPM for 1 h at 4 • C. After removing the cells and solid particles by centrifugation at 16,350× g for 20 min, the supernatant was mixed with 4 volumes of cold ethanol. The sample was then centrifuged at 16,350× g for 20 min to precipitate γ-PGA. The ethanol was allowed to evaporate and the dried precipitate was dissolved in 1 mL of deionized water, followed by the addition of 2 µL of DNase (5 U/µL) and incubation at 37 • C for 1 h to degrade the bacterial DNA.
Extracted γ-PGA and a standard poly-L-γ-glutamic acid sodium salt (molecular weight ≥750 kd; Sigma, USA) were electrophoresed following the method of Mamberti, et al. [34]. Briefly, the samples were mixed with 1 µL loading dye (5 mg/mL Bromophenol Blue, 50% [v/v] Glycerol in TAE buffer) and allowed to electrophorese in 1.5% agarose gel in TAE buffer (40 mM Tris [pH 7.6], 20 mM Acetic acid, and 1 mM EDTA) at 50 V for 90 min. γ-PGA was visualized by staining the gel with 0.5% methylene blue in 3% acetic acid for 30 min followed by destaining in H 2 O.

Determination of Poly-γ-Glutamate Hydrolase (γ-PGA Hydrolase) Activity
In order to find out the source of the γ-PGA hydrolase enzyme, the bacterial culture supernatant (without bacteriophage infection), cytosolic contents of the bacteria (without bacteriophage infection), and bacteriophage-infected bacterial lysate were used for γ-PGA hydrolase enzyme assay. For bacterial culture supernatant preparation, B. subtilis ATCC 15245 were grown in TSB media at 37 • C in shaking condition (160 rpm) for 24 h followed by centrifugation at 26,200× g for 10 min. The cytosolic contents of the bacteria were prepared after lysis of the control bacteria (without bacteriophage infection) by lysozyme (at 37 • C for 1 h) and sonication (Vibra-Cell™ Ultrasonic Liquid Processors, Newtown, CT, USA) for 10 min, followed by centrifugation at 12,300× g for 5 min to remove the cell debris. The bacteriophage-infected bacterial lysate was prepared by the infection of bacteriophage BSP10 (MOI 1) to B. subtilis ATCC 15245 culture in TSB media followed by incubation at 37 • C for 24 h in shaking condition (160 RPM), and the supernatant was collected after centrifugation at 26,200× g for 10 min.
γ-PGA hydrolase activity was assayed following the method of Kimura and Itoh [13]. Briefly, 500 µL of the reaction mixture (1 mg of γ-PGA, 10 mM sodium phosphate [pH 7.5], 150 mM NaCl, and 200 µL supernatant) was incubated at 37 • C for 1 h. Then, 2 µL of DNase (5 U/µL) was added to each tube and incubated at 37 • C for 30 min to degrade the bacterial DNA. The samples were resolved in 1.5% agarose gel as stated above.

Isolation, Morphology, Host Range Analysis, and One
Step Growth Curve of Bacteriophage BSP10 Bacteriophage BSP10 was isolated from a fermented soybean brick, Meju, by using B. subtilis ATCC 15245 as a host. The bacteriophage showed a clear plaque in TA soft agar after 12 h of incubation at 37 • C. Bacteriophage BSP10 was found to be a member of the Myoviridae family with an isometric head (82 ± 8 nm long in diameter; n = 6) and a contractile tail (length, 180 ± 10 nm; width, 24 ± 1 nm, n = 6) ( Figure 1). Similar to the other SPO1-related phages, two parallel interconnected planes perpendicular to the tail were clearly observed upon contraction ( Figure 1A,B).

Complete Genome Sequencing and Functional Annotation of Bacteriophage BSP10
The genome of bacteriophage BSP10 is 153,767 bp long with 42.1% G+C content. A higher read coverage (~2 fold) was observed in the region between nucleotides 78,776-84,660 (5885 bp) using Geneious software, suggesting the presence of a long terminal repeat similar to phiNIT1 and Grass in the Nit1virus genus [25,35,36]. A total of 236 ORFs (185 on the forward strand and 51 on reverse strand) and 5 tRNAs were predicted ( Figure 3). Among them, 57 ORFs (24.2%) were functionally annotated using the BLASTP algorithm and they were categorized into 7 groups (Table S2), including structural proteins, DNA replication, transcription and repair, nucleotide metabolism, DNA packaging, introns and homing endonuclease, host lysis, and other functions. No genes for integrase and antibiotic resistance were identified.

Complete Genome Sequencing and Functional Annotation of Bacteriophage BSP10
The genome of bacteriophage BSP10 is 153,767 bp long with 42.1% G+C content. A higher read coverage (~2 fold) was observed in the region between nucleotides 78,776-84,660 (5885 bp) using Geneious software, suggesting the presence of a long terminal repeat similar to phiNIT1 and Grass in the Nit1virus genus [25,35,36]. A total of 236 ORFs (185 on the forward strand and 51 on reverse strand) and 5 tRNAs were predicted ( Figure 3). Among them, 57 ORFs (24.2%) were functionally annotated using the BLASTP algorithm and they were categorized into 7 groups (Table S2), including structural proteins, DNA replication, transcription and repair, nucleotide metabolism, DNA packaging, introns and homing endonuclease, host lysis, and other functions. No genes for integrase and antibiotic resistance were identified.

Complete Genome Sequencing and Functional Annotation of Bacteriophage BSP10
The genome of bacteriophage BSP10 is 153,767 bp long with 42.1% G+C content. A higher read coverage (~2 fold) was observed in the region between nucleotides 78,776-84,660 (5885 bp) using Geneious software, suggesting the presence of a long terminal repeat similar to phiNIT1 and Grass in the Nit1virus genus [25,35,36]. A total of 236 ORFs (185 on the forward strand and 51 on reverse strand) and 5 tRNAs were predicted ( Figure 3). Among them, 57 ORFs (24.2%) were functionally annotated using the BLASTP algorithm and they were categorized into 7 groups (Table S2), including structural proteins, DNA replication, transcription and repair, nucleotide metabolism, DNA packaging, introns and homing endonuclease, host lysis, and other functions. No genes for integrase and antibiotic resistance were identified.

Comparative Genomics and Phylogenetic Position
The bacteriophage BSP10 genome showed 87.30%, 84.48%, and 92.12% nucleotide sequence homology (BLASTN) with the ICTV-classified members of Nit1virus genus, phiNIT1, Grass, and SPG24, respectively, whereas no significant similarity was observed with the representative members of the genus Agatevirus, B4virus, Bastillevirus, Bc431virus, Cp51virus, or Wphvirus (Table 1). It also exhibited 75.42% and 83.9% proteomic homology with the bacteriophage phiNIT1 and Grass, respectively (the proteomic homology of bacteriophage BSP10 and SPG24 was not calculated as the genome of bacteriophage SPG24 was found under-annotated in the NCBI database), and less than 50% homology with the other bacteriophages (Table 1) in the CoreGenes analysis.
When the bacteriophage BSP10 genome was compared with the other ICTV-classified Bacillus bacteriophages using CLANS, it was clustered with the members of previously-proposed Bastillelike group bacteriophages [37] ( Figure S1). Genome comparison was conducted by using a dot plot matrix and progressiveMauve alignment which showed a similar genomic arrangement between bacteriophage BSP10 and the other members of the Nit1virus genus ( Figures S2 and S3). In addition, heat-plot analysis of the translated protein comparison of whole bacteriophage genomes exhibited high similarities among the members of the Nit1virus genus and bacteriophage BSP10 ( Figure S4). Similarly, the phylogenetic tree based on heat-plot ( Figure S4) and the selected protein markers such as major capsid protein, terminase large subunit, and metallophosphatases (Figure 4), also grouped bacteriophage BSP10 with the members of the genus Nit1virus.

Comparative Genomics and Phylogenetic Position
The bacteriophage BSP10 genome showed 87.30%, 84.48%, and 92.12% nucleotide sequence homology (BLASTN) with the ICTV-classified members of Nit1virus genus, phiNIT1, Grass, and SPG24, respectively, whereas no significant similarity was observed with the representative members of the genus Agatevirus, B4virus, Bastillevirus, Bc431virus, Cp51virus, or Wphvirus (Table 1). It also exhibited 75.42% and 83.9% proteomic homology with the bacteriophage phiNIT1 and Grass, respectively (the proteomic homology of bacteriophage BSP10 and SPG24 was not calculated as the genome of bacteriophage SPG24 was found under-annotated in the NCBI database), and less than 50% homology with the other bacteriophages (Table 1) in the CoreGenes analysis.
When the bacteriophage BSP10 genome was compared with the other ICTV-classified Bacillus bacteriophages using CLANS, it was clustered with the members of previously-proposed Bastille-like group bacteriophages [37] ( Figure S1). Genome comparison was conducted by using a dot plot matrix and progressiveMauve alignment which showed a similar genomic arrangement between bacteriophage BSP10 and the other members of the Nit1virus genus ( Figures S2 and S3). In addition, heat-plot analysis of the translated protein comparison of whole bacteriophage genomes exhibited high similarities among the members of the Nit1virus genus and bacteriophage BSP10 ( Figure S4). Similarly, the phylogenetic tree based on heat-plot ( Figure S4) and the selected protein markers such as major capsid protein, terminase large subunit, and metallophosphatases (Figure 4), also grouped bacteriophage BSP10 with the members of the genus Nit1virus.  BLASTP comparison of bacteriophage BSP10 ORFs showed that 75 (31.8%) and 48 (20.3%) of ORFs had a high resemblance (≥99%) with bacteriophage phiNIT1 and Grass, respectively. On the other hand, 25 and 32 ORFs of bacteriophage BSP10 were found to be unique (spanned throughout the genome) compared with phiNIT1 and Grass, respectively, while 16 ORFs of phiNIT1 and 33 ORFs of Grass are absent in bacteriophage BSP10 genome (again, as the genome of bacteriophage SPG24 was not annotated properly, the comparison was not made).

Identification of Poly-Gamma-Glutamate Hydrolase Homolog Gene in Bacteriophage BSP10 Genome
The bacteriophage BSP10 genome contained a poly-gamma-glutamate hydrolase (γ-PGA hydrolase) homolog (ORF172, 208 amino acids), which might function to breakdown γ-PGA, which is responsible for the sticky appearance as well as the functional properties of soybean-based fermented foods. It showed high similarities (BLASTP) with those of Grass (NCBI protein id. YP_008771398.1, 99%) and phiNIT1 (NCBI protein id. YP_008318292.1, 98%).

Growth Inhibition of B. subtilis ATCC 15245 and Growth of Bacteriophage BSP10 in Liquid Culture
Bacteriophage BSP10 effectively inhibited the growth of B. subtilis ATCC 15245 in liquid culture. Clear lysis was observed 3 h postinfection for MOI 1 (OD600 of 0.00) and 4 h postinfection for MOI 0.1 (0.00), whereas the OD600 of the control (no bacteriophage addition) was 1.32 ± 0.10 after 5 h (corresponding to 3 h postinfection in the bacteriophage-treated sample) and 1.42 ± 0.04 after 6 h (corresponding to 4 h postinfection) of bacterial inoculation ( Figure 5A).
After 24 h (48 and 72 h) incubation, the bacterial CFU counts were 8.6 ± 0.2 (8.0 ± 0.4 and 7.7 ± 0.1 log10 CFU/mL) and 3.1 ± 1.3 log10 CFU/mL (4.0 ± 0.6 and 3.9 ± 0.6 log10 CFU/mL) for the control (without bacteriophage) and the MOI 0.1 treated set, respectively. Interestingly, no bacterial regrowth was detected for MOI 1 after 24, 48, and 72 h. With regard to bacteriophage growth, PFU counts rapidly increased, and after 3 h postinfection reached 10.0 ± 0.2 log10 PFU/mL (MOI 1) and 9.8 ± 0.1 log10 PFU/mL (MOI 0.1) ( Figure 5B). No significant changes in the PFU counts were observed for the next 72 h. BLASTP comparison of bacteriophage BSP10 ORFs showed that 75 (31.8%) and 48 (20.3%) of ORFs had a high resemblance (≥99%) with bacteriophage phiNIT1 and Grass, respectively. On the other hand, 25 and 32 ORFs of bacteriophage BSP10 were found to be unique (spanned throughout the genome) compared with phiNIT1 and Grass, respectively, while 16 ORFs of phiNIT1 and 33 ORFs of Grass are absent in bacteriophage BSP10 genome (again, as the genome of bacteriophage SPG24 was not annotated properly, the comparison was not made).

Identification of Poly-Gamma-Glutamate Hydrolase Homolog Gene in Bacteriophage BSP10 Genome
The bacteriophage BSP10 genome contained a poly-gamma-glutamate hydrolase (γ-PGA hydrolase) homolog (ORF172, 208 amino acids), which might function to breakdown γ-PGA, which is responsible for the sticky appearance as well as the functional properties of soybean-based fermented foods. It showed high similarities (BLASTP) with those of Grass (NCBI protein id. YP_008771398.1, 99%) and phiNIT1 (NCBI protein id. YP_008318292.1, 98%).

Growth Inhibition of B. subtilis ATCC 15245 and Growth of Bacteriophage BSP10 in Liquid Culture
Bacteriophage BSP10 effectively inhibited the growth of B. subtilis ATCC 15245 in liquid culture. Clear lysis was observed 3 h postinfection for MOI 1 (OD 600 of 0.00) and 4 h postinfection for MOI 0.1 (0.00), whereas the OD 600 of the control (no bacteriophage addition) was 1.32 ± 0.10 after 5 h (corresponding to 3 h postinfection in the bacteriophage-treated sample) and 1.42 ± 0.04 after 6 h (corresponding to 4 h postinfection) of bacterial inoculation ( Figure 5A).

B. subtilis ATCC 15245 Growth Inhibition by Bacteriophage BSP10 in Cheonggukjang Fermentation
The fast-fermented soybean product Cheonggukjang was prepared to check the effect of bacteriophage BSP10 on B. subtilis growth inhibition during soybean fermentation. Bacteriophage BSP10 was artificially contaminated at 10 2 , 10 4 , and 10 6 PFU/g in autoclaved soybean which contained 10 4 CFU/g (which is the standard concentration used in factories) of B. subtilis ATCC 15245. After 24 h incubation, statistically significant differences (p < 0.05) were observed among all the treatments and the bacterial numbers were 8.85 ± 0.06, 8.34 ± 0.10, and 7.11 ± 0.13 log10 CFU/g for 10 2 , 10 4 , and 10 6 PFU/g addition of bacteriophage BSP10, respectively, as compared to 9.16 ± 0.04 log10 CFU/g of the control (no bacteriophage) ( Figure 6A). B. subtilis counts increased in all treatments after 48 h incubation, but the numbers were significantly lower in 10 4 PFU/g (8.56 ± 0.08 log10 CFU/g) and 10 6 PFU/g (8.29 ± 0.21 log10 CFU/g) of bacteriophage BSP10-treated sets as compared to control (9.51 ± 0.09 log10 CFU/g) and 10 2 PFU/g treated sets (9.31 ± 0.16 log10 CFU/g) ( Figure 6A). Similarly, after 72 h, bacterial numbers increased in all of the sets (except in 10 6 PFU/g treated flasks). No statistically significant difference in the bacterial numbers was observed between control (9.78 ± 0.09 log10 CFU/g) and 10 2 PFU/g bacteriophage BSP10-treated set (9.63 ± 0.18 log10 CFU/g), whereas the numbers were significantly different in 10 4 PFU/g (9.10 ± 0.14 log10 CFU/g) and 10 6 PFU/g (8.26 ± 0.23 log10 CFU/g) treated sets as compared to the control ( Figure 6A). propagation of bacteriophage BSP10 during Cheonggukjang fermentation (B). Cheonggukjang was prepared by B. subtilis ATCC 15245 (10 4 CFU/g) and artificial contamination of bacteriophage BSP10 at a concentration of 0 (control), 10 2 , 10 4 , and 10 6 PFU/g followed by incubation at 37 °C for a period of 72 h. At the end of every 24 h period, B. subtilis growth was monitored using TSA and bacteriophage counts was checked using TA soft agar, as described in Materials and Methods section. Data were presented as mean ± SD. Different superscript letters in each segment are significantly different (p < 0.05) according to ANOVA (Duncan's multiple range tests). ND: Not detected.

B. subtilis ATCC 15245 Growth Inhibition by Bacteriophage BSP10 in Cheonggukjang Fermentation
The fast-fermented soybean product Cheonggukjang was prepared to check the effect of bacteriophage BSP10 on B. subtilis growth inhibition during soybean fermentation. Bacteriophage BSP10 was artificially contaminated at 10 2 , 10 4 , and 10 6 PFU/g in autoclaved soybean which contained 10 4 CFU/g (which is the standard concentration used in factories) of B. subtilis ATCC 15245. After 24 h incubation, statistically significant differences (p < 0.05) were observed among all the treatments and the bacterial numbers were 8.85 ± 0.06, 8.34 ± 0.10, and 7.11 ± 0.13 log10 CFU/g for 10 2 , 10 4 , and 10 6 PFU/g addition of bacteriophage BSP10, respectively, as compared to 9.16 ± 0.04 log10 CFU/g of the control (no bacteriophage) ( Figure 6A). B. subtilis counts increased in all treatments after 48 h incubation, but the numbers were significantly lower in 10 4 PFU/g (8.56 ± 0.08 log10 CFU/g) and 10 6 PFU/g (8.29 ± 0.21 log10 CFU/g) of bacteriophage BSP10-treated sets as compared to control (9.51 ± 0.09 log10 CFU/g) and 10 2 PFU/g treated sets (9.31 ± 0.16 log10 CFU/g) ( Figure 6A). Similarly, after 72 h, bacterial numbers increased in all of the sets (except in 10 6 PFU/g treated flasks). No statistically significant difference in the bacterial numbers was observed between control (9.78 ± 0.09 log10 CFU/g) and 10 2 PFU/g bacteriophage BSP10-treated set (9.63 ± 0.18 log10 CFU/g), whereas the numbers were significantly different in 10 4 PFU/g (9.10 ± 0.14 log10 CFU/g) and 10 6 PFU/g (8.26 ± 0.23 log10 CFU/g) treated sets as compared to the control ( Figure 6A).  and propagation of bacteriophage BSP10 during Cheonggukjang fermentation (B). Cheonggukjang was prepared by B. subtilis ATCC 15245 (10 4 CFU/g) and artificial contamination of bacteriophage BSP10 at a concentration of 0 (control), 10 2 , 10 4 , and 10 6 PFU/g followed by incubation at 37 • C for a period of 72 h. At the end of every 24 h period, B. subtilis growth was monitored using TSA and bacteriophage counts was checked using TA soft agar, as described in Materials and Methods section. Data were presented as mean ± SD. Different superscript letters in each segment are significantly different (p < 0.05) according to ANOVA (Duncan's multiple range tests). ND: Not detected.

Degradation of γ-PGA in Cheonggukjang and γ-PGA Hydrolase Assay
To check the degradation of γ-PGA in Cheonggukjang, γ-PGA was extracted from the bacteriophage-treated and -untreated Cheonggukjang samples, and electrophoresed in 1.5% agarose gel. The fragmentation of γ-PGA (low molecular weight) was observed in the bacteriophage-treated Cheonggukjang samples, while a clear band of high molecular weight γ-PGA was noticed for the bacteriophage-untreated sample ( Figure 7A).
γ-PGA hydrolase activity was undetectable in the bacterial culture supernatant (no bacteriophage infection) and the bacterial cytosolic content. On the other hand, the degraded γ-PGA was observed in the bacteriophage-treated bacterial lysate ( Figure 7B).

Degradation of γ-PGA in Cheonggukjang and γ-PGA Hydrolase Assay
To check the degradation of γ-PGA in Cheonggukjang, γ-PGA was extracted from the bacteriophage-treated and -untreated Cheonggukjang samples, and electrophoresed in 1.5% agarose gel. The fragmentation of γ-PGA (low molecular weight) was observed in the bacteriophage-treated Cheonggukjang samples, while a clear band of high molecular weight γ-PGA was noticed for the bacteriophage-untreated sample ( Figure 7A).
γ-PGA hydrolase activity was undetectable in the bacterial culture supernatant (no bacteriophage infection) and the bacterial cytosolic content. On the other hand, the degraded γ-PGA was observed in the bacteriophage-treated bacterial lysate ( Figure 7B).

Discussion
As previously acknowledged, B. subtilis is used in different soybean-based fermented food industries, and is responsible for the fermentation and qualities of the final products [4,5]. Considering the similar roles of lactic acid bacterial bacteriophages in the dairy industry, it can be assumed that contamination by a B. subtilis-infecting bacteriophage might create a serious problem, such as the delay of fermentation process leading to economic losses and changes in the functional properties of the final product. The knowledge about the prevalence, diversity, classification, and properties of bacteriophages is essential for the fermentation industry to develop antiphage strategies

Discussion
As previously acknowledged, B. subtilis is used in different soybean-based fermented food industries, and is responsible for the fermentation and qualities of the final products [4,5]. Considering the similar roles of lactic acid bacterial bacteriophages in the dairy industry, it can be assumed that contamination by a B. subtilis-infecting bacteriophage might create a serious problem, such as the delay of fermentation process leading to economic losses and changes in the functional properties of the final product. The knowledge about the prevalence, diversity, classification, and properties of bacteriophages is essential for the fermentation industry to develop antiphage strategies and minimize the risk of bacteriophage infections [8]. The availability of sufficient genome sequences of lactic acid bacteria-infecting bacteriophages has helped in the establishment of the taxonomy and PCR-based detection tools for the lactic acid bacterial bacteriophage in dairy environments [9,10]. In contrast, relatively few B. subtilis-infecting bacteriophage whole genome sequences (24 sequences including bacteriophage BSP10) are available in the NCBI database until now.
In this study, we have isolated a Myoviridae family bacteriophage, BSP10 from a dried soybean-fermented brick, Meju (Figure 1). Bacteriophage BSP10 exhibited high sequence homologies and similar genomic arrangements (whole genome comparison using dot plot matrix and progressiveMauve alignment) with the members of the Nit1virus genus (phiNIT1, Grass, and SPG24) ( Figures S2 and S3). Moreover, in phylogenetic analysis (using Phylogeny.fr and SplitsTree programs), bacteriophage BSP10 was clustered with the members of the Nit1virus genus (Figure 4 and Figure S4). The above-mentioned findings clearly suggest that bacteriophage BSP10 belongs to the Nit1virus genus.
According to the current ICTV classification [25], 95% DNA sequence identity is the criterion for the demarcation of a new species in the genus Nit1virus. The whole genome sequence comparison (BLASTN) of bacteriophage BSP10 with phiNIT1 (87.30%), Grass (84.48%), and SPG24 (92.12%) ( Previously, it was proposed that the genes within the terminal redundancy region of the bacteriophage are the first to be expressed in the infected cell [40]. It was also suggested that their nucleotide sequence and the transcriptional units (e.g., promoters) might be optimized for efficient early expression in the host environment [40,41]. In this study, we found that the high read portion (suggesting the presence of long terminal repeat [35]) in the genome of bacteriophage BSP10 was relatively more diverse (84.39%, 71.04%, and 82.32% in BLASTN with the bacteriophages phiNIT1, Grass, and SPG24, respectively) than the other parts of the genomes (87.3%, 85.44%, and 92.12% with phiNIT1, Grass, and SPG24, respectively). Moreover, in this region, the bacteriophage BSP10 genome contains 4 unique ORFs which are absent in the genomes of bacteriophage phiNIT1 and Grass (annotation of SPG24 is incomplete). While a detailed study is needed to find out the role of the individual genes, it can be hypothesized that the genes in the redundancy region (bacteriophage BSP10 genome) might be naturally selected or evolved for efficient early expression in the host environment which endows diversity and even uniqueness to the bacteriophage.
Nit1virus bacteriophages were found in a wide range of environments (phiNIT1 and BSP10, soybean-based fermented foods; Grass, soil samples; SPG24, decayed rice hay manure as mentioned in the Genbank file) and different geographical locations (phiNIT1, Japan; Grass, USA; SPG24 and BSP10, Korea) [13,36]. Considering the survivability in different environmental conditions of Nit1virus bacteriophages, it can be postulated that these widespread bacteriophages might be a threat to soybean-based fermented food industries and other industries where B. subtilis is used as a fermenting bacteria.
Along with the genome sequence analysis, morphological features are also important to classify the bacteriophage. The overall morphological features of bacteriophage BSP10 were similar to that of phiNIT1 (TEM images of other members in the Nit1virus genus are not available), possessing a long, contractile tail and an isometric head; baseplates moved upwards and the tail tube was extended from the baseplates and tail sheath (Figure 1). On the other hand, the dimensions of bacteriophage BSP10 (Figure 1) were found to be smaller (head diameter, 82 nm; and tail length, 180 nm) than phiNIT1 (head diameter, 100 nm; and tail length, 250 nm) [42]. Morphological variations between species of the same genus were also observed in the Bc431virus genus; the tail length of the species Bacillus virus JBP901 was 170 ± 5 nm [6], whereas that of another species Bacillus virus BCP82 was 210 nm on average [15]. The importance of this size difference among the different species in the Nit1virus genus is currently not clear.
Many studies explored the beneficial uses of bacteriophage to control pathogenic bacteria such as Listeria monocytogenes, Escherichia coli, Salmonella spp., Campylobacter spp., Enterobacter sakazakii, and B. cereus [15,[45][46][47]. In contrast, the detrimental effects of the bacteriophage (lactic acid bacterial bacteriophages could delay or abort dairy fermentation by constant lysis of fermentative bacteria) were also reported [7,8]. We observed that bacteriophage BSP10 efficiently inhibited the growth of B. subtilis ATCC 15245 in liquid culture ( Figure 5A) as well as in Cheonggukjang ( Figure 6A). Significant reductions in bacterial numbers were observed (up to 112 folds) for 10 4 and 10 6 PFU/g treated sets throughout the Cheonggukjang fermentation ( Figure 6A). Even 10 2 PFU/g of bacteriophage BSP10 decreased bacterial numbers (2 fold) after 24 h ( Figure 6A). Hence, it is clear that a very low number of bacteriophage can hamper bacterial growth and could delay or abort the soybean fermentation process.
In spite of the presence of high numbers of bacteriophage after 24 h, the bacterial counts gradually increased in all the sets at 48 and 72 h incubation, and no significant differences in bacterial numbers were observed for 10 2 PFU/g of bacteriophage BSP10-treated sets and the control at 48 and 72 h ( Figure 6A). It can be articulated that the incomplete inhibition of B. subtilis might be due to either the solid-state growth media where all of the bacteria may not be accessed by bacteriophage or the development of bacterial resistance. Foschino et al. [48] also stated a more effective growth inhibition of Lactobacillus sanfranciscensis by bacteriophage in liquid sourdough fermentation than the solid-state fermentation.
High bacteriophage titer was observed during the liquid culture inhibition ( Figure 5B) and Cheonggukjang fermentation ( Figure 6B), possibly due to the large burst size of bacteriophage BSP10 (~185, Figure 2) and/or constant presence of an adequate number of the host in the medium [49,50]. Although no reports have described the growth of B. subtilis-infecting bacteriophage during solid-state soybean fermentation, our findings might be comparable to the high titers of lactic acid bacterial bacteriophage in whey samples (10 9 -10 10 PFU/mL) collected after cheese fermentation [51].
γ-PGA, a sticky material, is produced in soybean-based fermented foods by bacterial (especially B. subtilis) fermentation [52]. It has several health-beneficial properties such as total cholesterol reduction, immune function enhancement, increased calcium absorption, and antitumor effects [53][54][55]. On the other hand, it was assumed that it might protect the bacteria from bacteriophage infection [13]. A number of bacteriophage genomes were reported to encode a putative γ-PGA-degrading enzyme, and recently, a purified γ-PGA hydrolase from phiNIT1 was shown to be able to actually breakdown γ-PGA [34,42]. However, to the best of our knowledge, whether or not B. subtilis-infecting bacteriophage directly affects γ-PGA qualities during fermentation in the soybean medium has not been experimentally proven.
In this study, we found the fragmentation of γ-PGA in the bacteriophage BSP10-treated samples during Cheonggukjang fermentation ( Figure 7A). Furthermore, we proved that γ-PGA hydrolase activity was not detected in the bacterial culture supernatant (no bacteriophage infection) or bacterial cytosolic contents (no bacteriophage infection) but in the bacteriophage-infected bacterial lysate ( Figure 7B). Taken together, our data strongly suggest that γ-PGA hydrolase activity is responsible for the degradation of γ-PGA and the enzyme (likely to be encoded by ORF172) is originated from the bacteriophage BSP10.
The production of a low molecular weight γ-PGA in bacteriophage-treated samples could negatively affect the functionalities of γ-PGA. Taken together with the reports on the detrimental effects of lactic acid bacterial bacteriophage in dairy environments, such as reduced acidification of milk in presence of lactic acid bacterial bacteriophage [7,56], our findings suggest that the contamination of B. subtilis-infecting bacteriophage in Cheonggukjang fermentation not only retards bacterial growth but also significantly lowers product qualities.

Conclusions
A Myoviridae family B. subtilis-infecting bacteriophage BSP10 was isolated from Meju and found to be a new species in the Nit1virus genus. We proved that bacteriophage BSP10 infection in Cheonggukjang fermentation not only affected bacterial growth but also decreased the quality of the soybean-based fermented food by degrading the functional compound, γ-PGA. This study will provide useful information for future research on the host-bacteriophage interaction and help to establish a clear classification scheme of B. subtilis-infecting bacteriophage.
Supplementary Materials: The following are available online at http://www.mdpi.com/1999-4915/10/5/240/s1, Figure S1: CLANS analysis of a total of 10 ICTV (International Committee on Taxonomy of Viruses)-classified Bacillus bacteriophages (phiNIT1, Grass, SPG24, phiAGATE, B4, Bastille, vB_BceM_Bc431v3, CP-51, W.Ph., and G), and bacteriophage BSP10. Edge weights were calculated from the P values of the BLASTN high-scoring segment pairs (e-value cut-off = 1e −5 ). The network was visualized after 25,000 runs using the CLANS software package; Figure S2: Whole-genome nucleotide dot plot of bacteriophage BSP10 and other ICTV (International Committee on Taxonomy of Viruses)-classified Bacillus bacteriophage (phiNIT1, Grass, SPG24, phiAGATE, B4, Bastille, vB_BceM_Bc431v3, CP-51, W.Ph., and G). Dot plots of whole genome sequences were generated using Gepard program; Figure S3: Bacteriophage BSP10 genome is compared with genus Nit1virus (bacteriophage phiNIT1, Grass, and SPG24), Agatevirus (phiAGATE), B4virus (B4), Bastillevirus (Bastille), Bc431virus (vB_BceM_Bc431v3), Cp51virus (CP-51), and Wphvirus (W.Ph.) as well as unassigned phage G using progressiveMauve alignment. Colored blocks indicate 1 to 1 best alignment and the sequence similarity is indicated by similarity plot within colored blocks; Figure S4: Heat-plot showing the results of genome comparison of bacteriophage BSP10 with other ICTV-classified Bacillus phage and the resulting phylogenetic tree. The similarity values were calculated using Gegenees software based on pairwise translated comparison of the analyzed sequences (tBLASTx method, fragment size −50, step size −25, threshold 10%). The heat plot colors reflect this similarity, ranging from low (red) to high (green). The tree was constructed with SplitsTree using the neighbor-joining method. The scale bar represents a 10% difference in the average tBLASTx score; Table S1: Sensitivity of Bacillus subtilis strains against Bacillus bacteriophage BSP10; Table S2: Functional grouping of annotated ORFs in Bacillus bacteriophage BSP10; Table S3: Homing endonuclease in bacteriophage BSP10 genome predicted by HHpred program.
Author Contributions: K.G., A.S., and K.P.K. conceived and designed the experiments; K.G. and H.S.K. performed the experiments; W.B.H. and J.E.K. performed the one-step growth assay of BSP10; K.G., A.S., and K.P.K. analyzed the data; K.G. and K.P.K. wrote the paper; all authors read and approved the final manuscript.