Original papers
A Survey of Phage Contamination in Natto-producing Factories and Development of Phage-resistant Bacillus subtilis (natto) Strains
2018 Volume 24 Issue 3 Pages 485-492
Details
2018 Volume 24 Issue 3 Pages 485-492
Eight natto-producing factories were surveyed for the contamination of phages that infect Bacillus subtilis (natto). A total of 160 sampling points, including manufacturing pipelines, instruments, floors, drainage ports, containers, and building facilities were inspected by wiping the surfaces. Virulent phages were detected from 28 survey points in four of the factories. Phages were frequently found in factories that had experienced product spoilage by phage contamination within the past 3 years. Isolated phages were not clonal and variations in their genomes were observed, although some phages were very similar. We developed phage-resistant B. subtilis (natto) strains using isolated phages as indicators. They were resistant against all phages found in the survey and those isolated decades ago from spoiled natto. The yueB gene that codes the receptor for phage SPP1 was not essential for the phages tested in this study to infect B. subtilis (natto) strains.
Phage contamination is a concern of manufacturers who produce fermented foods, and the development of resistant strain(s) is important to overcome this problem (Samson and Moineau, 2013; Brüssow, 2001). Bacillus subtilis (natto) is the bacterium which is used to produce natto, a fermented soybean food (Steinkraus, 2004; Kubo et al., 2011; Kamada et al., 2015, 2014; Nishito et al., 2010). Phages that infect B. subtilis (natto) spoil the natto (Kimura and Itoh, 2003; Nagai and Yamasaki, 2009). Product spoilage has prompted a contamination survey of natto factories, but reports on systematic and quantitative analyses of phage contamination in natto production facilities are limited. In addition, the development of natto-fermenting strains that are resistant against a variety of phages has not been performed.
In this study, we surveyed eight natto-producing factories, including some that had no experience of product spoilage during the past 3 years. Areas and items in manufacturing facilities such as containers, carriers, floors, and drainage ports were examined for phage contamination. Phages found were subjected to further study for classification and used to cultivate phage-resistant strains.
Major B. subtilis phages can be morphologically classified into one of three groups: a contractile tail Myoviridae, a long, noncontractile tail Siphoviridae, or a short tail Podoviridae, and they have been used for species phage typing (Ackermann, 2011; Ackermann et al., 1995). To date, the phages found in spoiled natto belong to the Myoviridae family and can be classified into two groups (group I and II) based on the similarity of their genomes (Nagai and Yamasaki, 2009). Among them the entire genome information of ϕNIT1 phage of the group II is available (Ozaki et al., 2016).
It is known that some Bacillus phages have pghP and levP genes that encode enzymes that break down extracellular sugar and amino acid polymers, fructooligosaccharide (levan) and poly-γ-glutamic acid (γPGA), respectively, to facilitate infection and propagate phage progenies (Kimura and Itoh, 2003; Ozaki et al., 2016). The appearance of pghP and levP in the genome of the contaminated phage was also examined.
Phages recognize receptor molecule (s) located on the cell surfaces; this results in a successful infection. The SPP1 phage requires the YueB receptor protein to infect B. subtilis cells (Baptista et al., 2008; Sao-Jose et al., 2006; Vinga et al., 2012). Receptors for other B. subtilis phages are largely unknown. We created a yueB knockout mutant to assess the involvement of YueB in phage contamination in natto factories.
B. subtilis and phage strains Three natto-fermenting B. subtilis strains and a total of 56 environmental B. subtilis strains including the laboratory strain were employed in this study. Natto-fermenting B. subtilis strains, Namegata-2-2, Miyagi-4, Miyagi-4100, and a phage-resistant B. subtilis strain Bandou-21-2 were obtained as previously reported (Kubo et al., 2011; Kubo et al., 2013). B. subtilis NFRI8322 was isolated from Tua Nao (a natto-like food of Chiang Rai, Thailand) (Inatsu et al., 2002; Kamada et al., 2015). A commercial natto-fermenting strain Miyagino and ϕNIT1 phage were used as reference strains (Kimura and Itoh, 2003; Nishito et al., 2010; Ozaki et al., 2016). Phages JNDMP and ONPC previously isolated from spoiled natto were obtained from the Genebank of the National Agriculture and Food Research Organization (i). B. subtilis phage no. 8 was isolated from spoiled natto on Iki Island, Nagasaki prefecture. B. subtilis (natto) Miyagino was used as a host to amplify B. subtilis (natto) phages. For phage typing experiments, a total of 49 B. subtilis strains that belong to 29 different phage types and SPP1 phage were obtained from Dr. Ackermann (Ackermann et al., 1995). The laboratory strain B. subtilis 168 was kindly provided by Dr. F. Kawamura and used as a host to amplify SPP1 phage.
Survey of phage contamination Eight natto factories located in the Mito area, Hitachioota area, and Hitachioomiya area, Ibaraki prefecture were surveyed for phage contamination between June and August in 2012. Surface of tested objects (10 cm × 10 cm) were swiped with sterilized swabs. The swabs were washed in 1 mL SM buffer [100 mM NaCl, 8 mM MgSO4·7H2O, 50 mM Tris, 0.01% (w/v) gelatin, pH 7.5], and phages extracted to the buffer were detected by a plaque formation assay on soft agar [0.7% (w/v)] using the B. subtilis Miyagino strain as an indicator, as previously described (Nagai and Yamasaki, 2009). Isolated phages were classified as group I or II, as previously described (Nagai and Yamasaki, 2009).
For each sampling source correlation between the product spoilage due to phage contamination in the past three years and the detection of phages in this survey was evaluated by Fisher's exact test. Empirical p-values were obtained by using the arules package developed for the R statistical environment (Hahsler et al., 2005; R Development Core Team, 2010).
Cultivation of phage-resistant strains Natto-fermenting strains were transformed with the genomic DNA of a natto-non-fermenting strain that is resistant to ϕNIT1 to develop phage-resistant natto starter strains. Bacterial cells of natto-fermenting strains, Namegata-2-2 and Miyagi-4100, were grown in 2 mL SPII medium (Harwood and Cutting, 1990) to the middle log phase at 37°C, and genomic DNA (1 µg) of phage-resistant Bandou-21-2 was added to the cell suspension. The mixture was further incubated for 30 min, and then ϕNIT1 phage (final 2 × 104 Pfu/mL) was added and incubated for 1 h. Cells were spread on LB agar plates and colonies that appeared were picked up and inoculated into 2 mL LB medium containing 2 × 104 Pfu/mL of ϕNIT1. After 24 h incubation at 37°C with shaking, the surviving cells were screened again for resistance to ϕNIT1 by plaque forming assay. Candidate strains were further screened for their applicability to natto production and resistance against phages obtained in the survey.
Knockout of the yueB gene The yueB gene (3.2 kb) of the B. subtilis (natto) Miyagino strain was amplified by polymerase chain reaction with primer pairs (Fw BS3-54, 5′-ATG ACA GAA CAA CGA AAA AGC TTG-3′, and Rv BS3-55, 5′-GCT TCA TAC GTT TCA TCG CTT TC-3′) using DNA polymerase (KOD-Plus, Toyobo, Osaka Japan). The amplified DNA fragment was inserted into the HincII site of pUC118 to construct pNAG575. The 1.5 kb Ermr cassette fragment excised from the pDG646 plasmid (Guérout-Fleury et al., 1995) by digestion with BamHI was inserted to the BglII site within the yueB of pNAG575, creating pNAG576. pNAG576 was linearized with Eco0109I digestion and used to disrupt the yueB gene of the Miyagino strain by homologous recombination, as previously described (Do et al., 2011). Correct disruption of yueB with the Ermr cassette was verified by Southern blot hybridization.
DNA manipulation and sequencing Escherichia coli DH5α (Toyobo) was used for gene cloning and plasmid construction. The genomic DNA of B. subtilis cells and phages were purified, as described (Harwood and Cutting, 1990). DNA sequencing analyses were performed with sequencing BigDye®Terminator V3.1 cycle sequencing kit and ABI Prism 310 genetic analyzer (Applied Biosystems, MA USA) following the manufacturer's instructions. Oligo nucleotide primers for sequencing were appropriately synthesized (Tsukuba Oligo service, Tsukuba, Japan).
Southern blot analysis Genomic DNA was digested with the indicated restriction endonuclease and fractionated on 1% agarose gel in TAE buffer (40 mM Tris-base, 20 mM acetic acid, 1 mM EDTA), and then transferred to Hybond N+ membrane (Amersham Pharmacia Biotech). The membrane was probed with 10 ng/mL of a labeled DNA fragment using the Alkphos Direct Labeling reagent (GE Healthcare, USA). After hybridization, the hybridized signals were detected by a CDP-star detection reagent (GE Healthcare) and X-ray film (Fujifilm, Japan). The oligo nucleotide primer pairs: BS2-2, 5′-CATATGGCACAAACAGACACATATCCA-3′ and BS2-3, 5′-CTCGAGGCCATAATACTCTGCCTCTGCTTC-3′, and BS3-47, 5′-CTCGAGGTCCGCCTTCTTAATAGACAGA CT-3′ and BS3-48, 5′-CATATGAGCAAGTTTAGACCGAAA TTC-3′, were used to amplify the probe DNA fragments containing pghP and levP, respectively.
Natto production Natto was made as previously described (Kubo, et al., 2011). Briefly, soybeans were soaked in 20°C water for 16 h, and then steamed at 0.18 MPa for 30 min. The steamed soybeans were inoculated with spores (104 cfu/g) and placed on polystyrene paper packages (50 g/package). Natto fermentation was continued for 18 h at 39°C under 90% relative humidity (RH) followed by a further 2 h incubation at 20°C under 50% RH. Natto products were matured at 5°C for 24 h. We evaluated the produced natto for slimy texture, smell, and appearance as well as sensory score, as previously described (Kubo et al., 2011; Kubo et al., 2013).
Chemicals not mentioned were purchased from Wako Pure Chemicals (Osaka, Japan).
Phage contamination survey A summary of the phage contamination survey is shown in Table 1. A total of 160 points in eight natto factories in Ibaraki prefecture were inspected. Among them, phage contamination was mainly found in those that had experienced product spoilage within the past 3 years (factories F1, F2, and F5 in Table 1). On the contrary, only one survey point (factory F7, carrier) was found to be contaminated among the four natto factories that had no product spoilage within the 3 years. Heavy contamination was observed in factories F1 and F2 (Table 1). The machines used to manufacture natto were contaminated in these two factories, which implied that phages multiplied there and spread over the facility, including the containers, floor, and building (wall and window frames). An instrument of F5 and a carrier of F7 were contaminated, but phages were not detected in their manufacturing machines. This may explain the low-level contamination of F5 and F7. The factory scale and the phage contamination seemed to have no direct relationship (Table 1).
| Factory ID (production scale)* | Product spoilage in the past 3 years | Sampling source | Sampling point | Contamination found | Isolated phages** |
|---|---|---|---|---|---|
| F1 | yes | manufacturing machine | 6 | 3 | #13 |
| (100–200) | Platform, rack or container | 4 | 4 | #16, #17 | |
| Instrumentation | 3 | 3 | #14, #15 | ||
| floor in the factory | 1 | 1 | #18 | ||
| Carrier | 1 | 1 | |||
| building facility | 4 | 3 | |||
| F2 | yes | manufacturing machine | 8 | 1 | |
| (50–100) | Platform, rack or container | 13 | 6 | #7, #10, #11 | |
| instrumentation | 9 | 2 | |||
| drainage port | 1 | 0 | |||
| building facility | 7 | 2 | #6 | ||
| F3 | no | manufacturing machine | 9 | 0 | |
| (200<) | Platform, rack or container | 4 | 0 | ||
| instrumentation | 2 | 0 | |||
| floor in the factory | 1 | 0 | |||
| carrier | 2 | 0 | |||
| F4 | no | manufacturing machine | 10 | 0 | |
| (100–200) | instrumentation | 2 | 0 | ||
| floor in the factory | 5 | 0 | |||
| F5 | yes | manufacturing machine | 4 | 0 | |
| (200<) | Platform, rack or container | 2 | 0 | ||
| instrumentation | 9 | 1 | #5 | ||
| floor in the factory | 3 | 0 | |||
| drainage port | 1 | 0 | |||
| F6 | no | manufacturing machine | 5 | 0 | |
| (50–100) | Platform, rack or container | 1 | 0 | ||
| instrumentation | 1 | 0 | |||
| floor in the factory | 1 | 0 | |||
| drainage port | 1 | 0 | |||
| building facility | 4 | 0 | |||
| F7 | no | manufacturing machine | 5 | 0 | |
| (200<) | Platform, rack or container | 2 | 0 | ||
| instrumentation | 4 | 0 | |||
| floor in the factory | 2 | 0 | |||
| carrier | 2 | 1 | #9 | ||
| drainage port | 1 | 0 | |||
| F8 | yes | manufacturing machine | 11 | 0 | |
| (<50) | Platform, rack or container | 1 | 0 | ||
| instrumentation | 2 | 0 | |||
| floor in the factory | 1 | 0 | |||
| building facility | 5 | 0 | |||
| Total | 160 | 28 |
No phage contamination was found in F8, although it had experienced product contamination within the past 3 years (Table 1). Local environment of the factory may have affected the survival of phages. F8 manufactures processed foods made of natto and pickles, which was unique among the factories. In factories F1 and F2 besides normal natto, mildly dried natto is produced once the fermentation process is complete, which might have given the phages a chance to survive.
Correlation between the product spoilage in the past three years and the phage detection in this survey was further examined statistically by Fisher's exact test for each sampling source. For manufacturing machine and instrumentation which have routine contact with natto product p-values were 0.112 and 0.150, respectively. For other sampling sources which has less chance to contact natto products, p-values were 0.026 (platform, rack or container) and 0.012 (floor, carrier, building facilities, and drainage port). This may suggest that manufacturing pipelines are washed at intervals while phages could survive for several years on the surface of factory facilities. Aeolian dust could carry bacteria including Bacillus species (Maki et al., 2014). Likewise, phages might have been airborne.
Genome variations of isolated phages Genomic DNA of 18 phages as listed in Table 2 (No.1 to No. 18) was digested with EcoRV restriction enzyme, and the banding patterns were compared (Fig. 1). The restriction fragment patterns resembled each other, but were not homogeneous (Fig. 1). Phages isolated from the same factory demonstrated variations, and no phages had an identical EcoRV restriction profile to that of ϕNIT1, which was isolated from a spoiled natto product in the early 2000s (Kimura and Itoh, 2003). This may reflect diversity of natto-contaminating phages and/or a raid differentiation of the phage genome. Four natto-fermenting cultivars, Miyagino, Miyagi-4, Miyagi-4100, and Namegata-2-2 were sensitive to all 18 phages (Table 2). The laboratory strain B. subtilis 168 that was used as a host strain of SPP1 phage was resistant to all phages.
| phages | B. subtilis (natto) strains | |||||||
|---|---|---|---|---|---|---|---|---|
| Phage No. | source or name | Miyagino* | Miyagi-4* | Miyagi-4100* | Namegata-2-2* | Bandou-21-2 | Phr-4* | Phr-20* |
| 1 | F5 natto product | s | s | s | s | R | R | R |
| 2 | F5 natto product | s | s | s | s | R | R | R |
| 3 | F5 natto product | s | s | s | s | R | R | R |
| 4 | F5 natto product | s | s | s | s | R | R | R |
| 5 | F5 instrumentation | s | s | s | s | R | R | R |
| 6 | F2 building facility | s | s | s | s | R | R | R |
| 7 | F2 platform, rack or container | s | s | s | s | R | R | R |
| 8 | spoiled natto product from Iki island** | s | s | s | s | R | R | R |
| 9 | F7 container | s | s | s | s | R | R | R |
| 10 | F2 platform, rack or container | s | s | s | s | R | R | R |
| 11 | F2 platform, rack or container | s | s | s | s | R | R | R |
| 12 | ϕNIT1 | s | s | s | s | R | R | R |
| 13 | F1 manufacturing machine | s | s | s | s | R | R | R |
| 14 | F1 instrumentation | s | s | s | s | R | R | R |
| 15 | F1 instrumentation | s | s | s | s | R | R | R |
| 16 | F1 platform, rack or container | s | s | s | s | R | R | R |
| 17 | F1 platform, rack or container | s | s | s | s | R | R | R |
| 18 | F1 floor in the factory | s | s | s | s | R | R | R |
| 19 | SPP1 | R | R | R | R | R | NT | NT |
| 20 | JNDMP | s | s | s | s | R | R | R |
| 21 | ONPC | s | s | s | s | R | R | R |
R, resistant, s, sensitive, NT, not tested. The phage number and sources correspond to those in Table 1 and Fig. 1.
The banding pattern of EcoRV-digested fragments of phage genomic DNA. Phage genomic DNA was digested with EcoRV and subjected to 1% agarose gel electrophoresis. Lane numbers 1–18 correspond to the phage number listed in Table 1. F1 to F7 indicate factory IDs as described in Table 1. Iki (lane 8) indicates the phage isolated from Iki Island, Nagasaki. ϕNIT1 phage was loaded as a reference (lane 12).
Some B. subtilis strains produce an extracellular slimy glutamic acid polymer (poly-γ-glutamic acid, γPGA) and fructose polymer (levan). These compounds surrounding the cells are assumed to work as a barrier against phage attack and progeny expansion (Kimura and Itoh, 2003; Ozaki et al., 2016). On the contrary, phages appear to have evolutionally acquired enzymatic counter levers, hydrolases for the γPGA and the levan (Ozaki et al., 2016). We examined the presence of the γPGA hydrolase gene (pghP) and levanase gene (levP) among isolated phages (Fig. 2). pghP was found in all tested phages (100%, 18/18) and levP was found in 66.7% (12/18). The frequent lack of levP in the phage genome implied that γPGA had a more important contribution to host–phage competition than levan.
Appearance of the poly-γ-glutamate hydrolase gene (pghP, upper panel) and levanase gene (levP, lower panel) in phage genome. Phage genomic DNA was digested with EcoRV and analyzed by Southern hybridization. Labeled DNA fragments containing pghP and levP of the ϕNIT1 phage were used as probes to detect these genes, as described in the Materials and Methods.
Cultivation of the phage-resistant B. subtilis (natto) strain Previously, a phylogenetic study of natto-fermenting strains was performed using a collection of B. subtilis environmental strains (Kubo et al., 2011). None of the natto-applicable strains were resistant to phage ϕNIT1, and resistant strains were not natto-applicable (Kubo et al., 2011). Therefore, a natto non-fermenting strain (Bandou-21-2) that is resistant to ϕNIT1 phage was selected as a donor of genomic DNA for transformation to create phage-resistant natto-fermenting strains. Although the strain Bandou-21-2 was a natto non-fermenting strain, it is γPGA-positive, biotin auxotrophic, and has a swarming ability. These phenotypes are commonly found in natto-fermenting strains (Do et al., 2011; Kimura et al., 2011, 2009; Kubo et al., 2011). Among the phage-resistant strains collected, Bandou-21-2 appeared to be more closely related to natto-fermenting strains than the others.
Cells of B. subtilis (natto) strains, Namegata-2-2 and Miyagi-4100, were transformed with genomic DNA of Bandou-21-2 and exposed to a high dose of ϕNIT1 phage (2 × 104 pfu/mL). Colonies formed by the surviving cells were picked up and again screened for phage resistance. Candidate lineages were further screened by their applicability for natto production by evaluating natto made with them, as previously described (Kubo et al., 2013; Kubo et al., 2011). We obtained two natto-fermenting strains with resistance to ϕNIT1 phage, Phr-4, and Phr-20, from Namegata-2-2 and Miyagi-4100, respectively.
Phage-resistance spectrum of B. subtilis (natto) Phr-4 and Phr-20 Eighteen phages listed in Fig. 1 and phages JNDMP and ONPC that were previously isolated from spoiled natto independently (Nagai and Yamasaki, 2009) were used to examine the phage-resistance spectrum of Phr-4 and Phr-20 strains (Table 2). Phages from spoiled natto products were classified as group I or II based on their genome similarity. In spoiled products, group II phages were more abundant than group I phages (Nagai and Yamasaki, 2009). JNDMP is a group I phage. ONPC and the 18 phages were group II phages (data not shown). The parental Miyagi-4100 and Namegata-2-2 were sensitive to all phages tested (Table 2). Phr-4, Phr-20, and Bandou-21-2 strains were resistant to all tested phages, including both group I and II (Table 2).
The phage-resistant mechanism of Phr-4 and phr-20 and genetic factor (s) required for it remain unknown. However, among the 49 B. subtilis strains that belong to 29 phage types (Ackermann, et al., 1995), at least two strains differently responded to the two groups: strain IP50 was sensitive to JNDMP (group I) and resistant to ONPC (group II), and strain HER1346 was resistant to JNDMP and sensitive to ONPC. In addition, strain NFRI8322 isolated from natto-like food (Tua Nao) of Thailand was sensitive to JNDMP (group I) and resistant to ONPC (group II). Perhaps more than one genetic factor was simultaneously transferred from Bandou-21-2 to Namegata-2-2 and Miyagi-4100. This requires further experimental elucidation.
B. subtilis has genetic competence to integrate DNA fragments into its genome by homologous recombination (Dubnau, 1999). Phage-resistant and phage-sensitive B. subtilis strains share their niches (Kubo et al., 2011). Thus, the horizontal transfer of genetic elements we observed in this study could potentially occur naturally in the environment. This meets the biological requirements of the Cartagena Protocol on Biosafety (ii) and the Japanese Authority of Genetically Modified Organisms (iii) for regulations exemption.
yueB gene of the resistant strain Bacteria screen themselves from phages by blocking uptake of the phage genome, restriction/modification, a clustered regularly interspaced short palindromic repeats system, defective phage production, hiding the phage receptor protein by a physical barrier, and losing the phage receptor protein (Hyman and Abedon, 2010; Nagai, 2014). Among them, the phage receptor plays a direct and essential role in the early stage of the infection. In B. subtilis, the yueB gene is the only known phage receptor gene that is essential for SPP1 to attach to cell surface and introduce its genomic DNA into the host cell (Baptista et al., 2008; Sao-Jose et al., 2006; Vinga et al., 2012). YueB has a unique structure; the central part of YueB that is supposed to be located outside the cell membrane is very diverse, while the transmembrane domains (N- and C-terminal parts) are conserved (Vinga et al., 2012). The rapid differentiation of the central part of YueB suggests that there has been evolutional competition between SPP1 phage and host cells and that some phages might recognize differentiated YueB. We compared yueB genes of Miyagino, Miyagi-4100, Namegata-2-2, and the phage-resistant Bandou-21-2.
The three phage-sensitive strains commonly had YueB with the same amino acid sequence. YueB of Bandou-21-2 was almost identical to those of the natto-fermenting strains, but it had three amino acid substitutions (A700E, G731D, A757T) located at the beginning of the C-terminal conserved region. The nucleotide sequence of the yueB gene of the cultivated phage-resistant strains, Phr-4 and Phr-20, was identical to that of the phage resistant Bandou-21-2 which was used as the DNA donor for transformation. To clarify the relationship between phage resistance and yueB gene variation, phage sensitivity of the yueB null mutant was scrutinized. A disruption of the yueB gene of the Miyagino strain had no effect on susceptibility for natto phages, ϕNIT1, JNDMP, and ONPC. The yueB gene of B. subtilis is probably not involved in phage contamination and propagation in natto factories.
A phage contamination survey of natto factories revealed that the degree of contamination differed between factories and that once a manufacturing machine was contaminated, phages could spread throughout the facility including the floor, wall, and windows. Phages isolated from the same factory were not identical and had different genome structures, suggesting a diversity of natto-contaminating phages and/or a raid differentiation of the phage genome. Routine survey of phage contamination by surface swiping regardless of the occurrence of product spoilage is recommended as along with general cleanliness of the factory, especially for factories that produce the dry-type natto.
Phage-resistant natto-fermenting strains (Phr-4 and Phr-20) were cultivated in this study. They are not genetically modified organisms and may serve natto manufacturers as backup strains in a case of heavy phage contamination.
Acknowledgments We are grateful for the cooperation of natto companies during the surveillance study. SS is a UNU-Kirin fellow. The phage no. 8 was kindly given by Hiromasa Hasegawa. This work was supported in part by a Grant from the Ministry of Education, Culture, Sports, Science and Technology, Tokyo, Japan (Tokubetsu dengen shozaiken kagaku gijyutsu sinkou).
