Isolation and characterization of bacteriophages from soil against food spoilage and foodborne pathogenic bacteria

Microbial food spoilage and foodborne disease are the main challenges in the food industry regarding food shelf life. Current preservation methods are frequently associated with changes in organoleptic characteristics and loss of nutrients. For this reason, bacteriophage offers an alternative natural method as a biocontrol agent that can reduce bacterial contamination in food without altering the organoleptic properties. This study was conducted to isolate and characterize bacteriophage from soil to control food spoilage bacteria, such as Bacillus cereus and Bacillus subtilis, and foodborne pathogenic bacteria, such as enterotoxigenic Escherichia coli (ETEC) and enterohemorrhagic E. coli (EHEC). Isolation was done by agar overlay assay method, and phages BC-S1, BS-S2, ETEC-S3, and EHEC-S4 were recovered. The host range of all isolated phages tended to be narrow and had high specificity towards the specific bacteria. The phage efficiency were measured where ETEC-S3 showed no effectivity against B. cereus and EHEC-S4 showed low efficiency against Enteropathogenic E. coli (EPEC). Morphology analysis was conducted for phage BC-S1 and ETEC-S3 with Transmission Electron Microscopy (TEM), and it is shown to belong to the Caudovirales order. Phages BC-S1 and BS-S2 significantly reduced the host bacteria when applied to the cooked rice and pasteurized milk samples with miMOI of 0.1. While phage ETEC-S3 at miMOI of 0.001 and phage EHEC-S4 at miMOI of 1 also showed a significant reduction when applied to chicken meat and lettuce samples at storage temperatures of 4 °C and 28 °C. The highest bacterial reduction of 100% was shown by phage BC-S1 on pasteurized milk samples and reduction up to 96.06% by phage ETEC-S3 on chicken meat samples at 28 °C incubation.

have pathogenicity that leads to foodborne diseases. One of the main pathogens involved in diarrheal disease, which can also lead to death, is Diarrheagenic Escherichia coli (DEC) which can be found in soil and water, such as ETEC and EHEC 5 .
Many risk factors associated with bacterial food contamination are often related to its processing, preparation, storage, and handling practices. Conventional food preservation methods, such as pasteurization, highpressure processing, irradiation, and chemical or biological agents, are commonly used to help improve food safety. However, those treatments are frequently associated with changes in organoleptic characteristics, loss of nutrients, also toxic-threatening side effects 6,7 . Additionally, even with the variety of methods available, foodborne outbreaks still occur relatively often 8 . For this reason, finding an alternative preservation method to control food spoilage is required.
One promising and safe technique that addresses several shortcomings is bacteriophage biocontrol. This method uses lytic bacteriophages to specifically target pathogenic bacteria and eliminate or significantly reduce their levels in food in order to enhance the safety of food products. Bacteriophages are viruses that lyse living bacterial hosts. This lytic potential has been exploited in attempts to design a more natural antimicrobial approach to control bacteria at the various stages of food production 9 . They are highly host-specific, safe to consume, relatively inexpensive, and do not alter the organoleptic properties of food 10 . They can be found almost everywhere where live bacteria exist, such as soil, offering the possibility to isolate them for therapeutic purposes. Hence, the use of bacteriophages as alternative natural preservation is very promising 6,11 . Based on these backgrounds, this study aimed to isolate and characterize bacteriophages from the soil in controlling food spoilage and foodborne pathogenic bacteria, such as B. cereus, B. subtilis, ETEC, and EHEC.
subtilis, ETEC, EHEC, EPEC, and Vibrio cholerae. The host range of isolated bacteriophages was showed in Table 1. Besides their capability to lyse its host cell, phage ETEC-S3 also showed lytic activity against B. cereus and phage EHEC-S4 showed lytic activity against EPEC. While phage BC-S1 and phage BS-S2 showed that they could only lysis their host itself. They performed high host specificity that could not attack other bacteria, even of which belonged to the same genus.
Efficiency of plating (EOP). All isolated phages showed activity only in infecting specific bacteria, their bacterial host. However, phage ETEC-S3 was also found to be inefficiently attack B. cereus with EOP lower than 0.001, while phage EHEC-S4 also performed low efficiency with EOP 0.001-0.2 against EPEC (Table 2).

Minimum inhibitory multiplicity of infection (miMOI).
The bacteriophage MOI was carried out on eight different concentrations from 10 2 to 10 -5 . Positive control showed only host bacteria without adding bac- Table 1. Bacteriophages host range. S1, soil 1; S2, soil 2; S3 , soil 3; S4, soil 4. www.nature.com/scientificreports/ teriophages, while negative control showed only bacteriophages without adding the host bacteria. Based on the result, the highest inhibition for phage BC-S1 ( Fig. 1) and phage BS-S2 (Fig. 2) were shown for MOI 0.1. While ETEC-S3, MOI 0.01 to the highest MOI 100 can inhibit ETEC with no growth in the first 6 h of incubation, followed by re-growth of ETEC. Therefore, miMOI of phage ETEC-S3 was 0.001, whereas the graphical showed no bacterial growth of ETEC (Fig. 3). miMOI of phage EHEC-S4 was 1, which could inhibit the growth of EHEC within 10 h of incubation. Bacterial growth decreased as the MOI increased (Fig. 4).

Morphology analysis.
Bacteriophage BC-S1 and phage ETEC-S3 were continued for morphology determination due to their higher activities using transmission Electron Microscopy (TEM) as shown in Fig. 5. Phage BC-S1 performed an icosahedral head with about 75 nm diameter and about 90 nm contractile tail. While phage ETEC-S3 performed an icosahedral head with about 65 nm diameter and about 100 nm contractile tail.
Bacteriophage application on food samples. Each bacteriophage was tested for its ability to reduce the number of specific pathogenic bacteria on food samples. Phages BC-S1 and BS-S2 were applied to cooked rice and pasteurized milk, while Phages ETEC-S3 and EHEC-S4 were applied to chicken meat and lettuce to see the effect of their application on different food matrices and surfaces. The incubation was carried out overnight at refrigerated storage temperature (4 °C) and room temperature (28 °C). The results were shown in Table 3, where all isolated bacteriophages showed a reduction in all samples and treatments. The highest reduction percentage by phages BC-S1 and BS-S2 was found on pasteurized milk samples, while phages ETEC-S3 and EHEC-S4 showed the highest reduction on chicken meat samples at 28 °C of storage temperature.  www.nature.com/scientificreports/

Discussions
Bacteriophages are the most abundant life form on earth and can be found in nearly every habitat, such as soil, water, and food. Soil near organic waste disposal was used as a sample because it is an ideal source to isolate bacteriophages since it contains high numbers of diverse bacteria. Thus, the presence of its bacteriophages is potential. There are an estimated 1.5 × 10 8 bacteriophages per gram of agricultural soil 12 . Four bacteriophages were successfully recovered, namely phage BC-S1 against B. cereus, phage BS-S2 against B. subtilis, phage ETEC-S3 against ETEC, and phage EHEC-S4 against EHEC. These phages could be considered lytic bacteriophages due to their clear plaque formation in the agar overlay assay. However, in areas where bacteriophages are absent, the bacteria grow to the stationary phase and form a confluent, opaque layer or "lawn" in the soft agar overlay 13 . Some required criteria for applying bacteriophage as a biocontrol agent are obligately lytic, non-transducing, and toxin gene-free to ensure safety 14 . Lytic bacteriophages replicate by attaching, injecting nucleic acid, and lysing the host cell to produce the phage progeny. The "newborn" bacteriophages are then ready to start another cycle by infecting another bacteria cell. While in the lysogenic cycle, phages integrate their nucleic acid into the chromosome of the host cell and replicate with it as a unit without destroying the cell 15 . Therefore, bacteriophages should be obligately lytic to reduce the potency of toxin gene transfer that can increase their virulence 16 .
Knowing the concentration of the isolated bacteriophages is essential. Bacteriophage titer is one of the factors that can affect their effectiveness in phage therapy applications. A high titer value indicates better phage stability. Bacteriophage applications for therapeutic use require a high titer of lytic phages (10 9 PFU/mL) 17 . In this study, the titer of isolated bacteriophages varied from 10 5 to 10 10 PFU/mL. It is recommended to refresh the bacteriophage regularly to maintain the phage stability in a high titer because long-time storage may cause the titer to decline 18 .
Characterization of bacteriophage by host range determination is also one factor in evaluating the isolated phage ability against different strains of host bacteria 11 . Phage ETEC-S3 was found to have the capability to infect B. cereus, and phage EHEC-S4 can infect EPEC, but both showed low efficiency. All the isolated bacteriophages performed a highly specific or narrow host range, including phage BC-S1 or BS-S2, since it showed high lytic activity only for the bacterial host, while to several pathogenic bacteria showed low efficacy and showed no lytic activity on the others, even more. Generally, newly isolated bacteriophages can only infect hosts with the same general receptor type as the isolated host 14 . The location of the cell receptors varies depending on the phage and the host. It can be located on the cell wall, flagella, pili, capsules, or surface protein in bacteria cells. Phage cannot bind to the host cell if the receptors are inaccessible or non-complementary to the phage receptor binding protein 19 . Other studies suggest that nearly all bacteriophages isolated using a single host strain of bacteria may be more likely to have a narrow rather than broad host range. In many cases, this narrow property is desirable because it usually has great specificity to the host itself, preventing the killing of other species of bacteria and leaving the rest of the bacterial host intact. This narrow host range phage is also essential for phage cocktail development, which broadens the host range for phage therapy. In this case, characterization of the individual phage of the cocktail is needed 14,20 .
The Efficiency of Plating (EOP) was used to define the effectiveness of bacteriophage against target bacteria. EOP value of 0.5-1.0 is ranked as high efficiency, an EOP value of 0.2 to 0.5 is categorized as medium efficiency, while an EOP value of 0.001 to 0.2 is classified as low efficiency, and EOP below 0.001 is inefficient 21 . According Table 3. Bacteriophage application onto a variety of food samples with different storage temperatures. Data were shown in mean ± standard error value. Different letters in each column indicated significant differences, α ≤ 0.05. "*"Indicate significant differences between control and bacteriophage treatment for each sample. www.nature.com/scientificreports/ to the result, phage EHEC-S4 performed higher efficiency than phage ETEC-S3. Phage ETEC-S3 was considered ineffective towards B. cereus, while phage EHEC-S4 performed low efficiency towards EPEC. Both isolated phages showed high specificity since they showed high activity only against bacterial hosts. Low EOP can be caused by the action of host resistance systems blocking the intracellular virus development or due to poor bacteriophage adsorption to the host cells 22 . MOI is determined as PFU/CFU ratio, which counts only adsorbed phages attached to and then infected bacteria 23 . The minimum MOI value needs to be examined to determine the potential effective concentration that can be used. In this study, all isolated phages showed different optimum MOI to inhibit or lysis each host completely. The minimum MOI value obtained for phage BC-S1 and BS-S2 were 0.1, the MOI of phage ETEC-S3 was 0.001, and the MOI of phage EHEC-S4 was 1. Phage ETEC-S3 showed the lowest MOI compared with the others, indicating that phage ETEC-S3 is considered more effective because a lower phage concentration is needed to reduce the number of bacteria. Effective MOI value is affected by environmental conditions such as the number of infecting phages, the number of target cells to attach, also how fast and how much time is allowed for attachment 12,23 . Phage BC-S1, BS-S2, and EHEC-S4 gave a better result in lytic activity using higher MOI numbers. Bacterial growth decreased as the MOI increased because higher use of MOI enhances the probability of phage particles infecting their host bacteria. A lower MOI value still showed the reduction though it could not completely inhibit the bacterial host growth. More concentration of bacteriophages means more cells can be lysed, yielding rapid lytic activity 24 . However, these phages differed from phage ETEC-S3, where MOI higher than 0.001 could not completely inhibit the growth of the cells. The bacterial reduction was shown in the first 6 h of incubation, and then the bacteria continued to re-grow. Phage concentration that is too high in terms of MOI may lead to bacterial lysis via the enzymatic action of the phage lysins once the phages are attached to their receptors, without new virions even accessing the cell. Hence, the productive infection of bacteriophages is stopped, generating no more viruses to invade the rest of the pathogenic population 25 .
The classification of bacteriophage depends on its nucleic acid type and its morphology. A Bacteriophage is composed of a head and a tail. The head or capsid is a protein shell that envelops the genetic material in an icosahedron shape. The tails generally have six tail fibers that vary in size and hold protein receptors to recognize the attachment sites on the surface of bacterial cell walls for attachment to specific host cells 26,27 . Tailed bacteriophages (Caudovirales) are divided into four families based on their tail shape (ICTV). Myoviridae has a long, rigid, and contractile tail with 80-485 nm length, and the average head diameter is 85 nm. Siphoviridae has a noncontractile long and flexible tail with a 79-535 nm length and an average head diameter of 55 nm. Podoviridae has a non-contractile short tail below 40 nm length and a 58 nm average head diameter 28 . Also, the newly created Ackermannviridae has up to four tail spike proteins and can infect a wide range of Gram-Negative bacteria 29 . In this study, isolated bacteriophage BC-S1 and ETEC-S3 morphology were analyzed using TEM. Based on the result, both phages showed an icosahedral head that is attached to a contractile tail. It can be assumed that both phages belong to the member of the Caudovirales order. Molecular analysis of the genome of the bacteriophage is required for further research to know the classification specifically and to ensure these phages do not contain any virulence-associated genes as well as antibiotic resistance genes 30 .
Since bacteriophages cause bacterial death, their potential use as food preservatives has become increasingly appealing. In this research, bacteriophage BC-S1 and BS-S2 were applied to cooked rice and pasteurized milk to control B. cereus and B. subtilis growth, and bacteriophage ETEC-S3 and EHEC-S4 were applied to chicken meat and fresh lettuce to control ETEC and EHEC growth. All isolated phages could significantly reduce their host bacteria population in every sample at all storage temperatures. Room temperature and low-temperature storage were selected because it was generally common temperature used to store food and beverage for a short period.
Overall, bacterial reduction in pasteurized milk samples was higher than in cooked rice, and bacterial reduction in chicken meat samples was higher than in lettuce. These reduction capabilities may be affected by the food sample matrices. Limited diffusion and contact of bacteria and phage were responsible for the low efficacy. Phage particles may be required to reduce bacterial contamination on moist food surfaces and in liquids compared to a drier food matrix because of the increased "mobility" of phages in the presence of moisture. The initial contact of bacteriophage and bacterium often occurs by diffusion and Brownian motion. Therefore, liquid samples such as pasteurized milk allowed greater diffusion of the phages rather than on the solid matrix such as cooked rice due to the restricted motions of the solvent molecule 28,31 .
Treatment for chicken meat showed a higher reduction than lettuce due to its natural juices. When chicken meat does not have space around it, the heat and moisture cannot escape, leaving the chicken to steam in its own juices. Phages in lettuce samples are likely unable to reach and invade bacteria, considering lettuce has a dry food matrix. The passive movement of phages across food surfaces is limited due to the lack of moisture 32 . On the other hand, at the same time and temperature, bacterial reduction in lettuce was higher compared with chicken meat in treatment using phage EHEC-S4. The situation is different on solid food with even surfaces like lettuce, where the total surface area and its ability to absorb liquid from the phage suspension are easier than uneven surface areas like chicken meat. Food with uneven and large surface area is difficult to treat with phage because phage distribution is physically limited to reach all bacterial targets. In addition, target bacteria may be embedded within the rather complex food matrix, shielding them from diffusing phage particles 33 .
Likewise, incubation at 28 °C showed higher bacterial reduction than at 4 °C of incubation. Temperature can also affect phage activity, where phages depend on the growth of bacterial hosts for their replication. B. cereus, B. subtilis, and E. coli grow in the mesophilic temperature range, with an optimum at about 37 °C. An optimal growth temperature of the bacterial host promotes a better replication of phage particles. In contrast, at a lower temperature, the rate of phage replication was considerably decreased or halted due to the lower growth rate of their hosts 34 .
All isolated bacteriophages in this study effectively reduced targeted food spoilage and foodborne pathogenic bacteria. However, further studies were required to determine their activities against other pathogens and their www.nature.com/scientificreports/ stability to various environmental conditions and characterize their genomic properties to ensure their safety if they want to be used as a preservation alternative.

Methods
Bacteriophages were isolated from soil, then enriched and purified. Characterization of the isolated bacteriophages, including titer, host range, the efficiency of plating (EOP), the minimum inhibitory multiplicity of infection (miMOI), and morphology were determined. The application of the isolated bacteriophages to food samples was also evaluated. Bacteriophage purification and enrichment. The isolated lytic bacteriophages were purified by stabbing the clear plaque gently using a sterile tip. The tip was then placed into 10 mL of LB Broth and pipetted up and down to release the bacteriophage particles. Bacteriophages were enriched by adding 250 µL of mid-log phase bacterial host culture into the LB Broth and incubated at 37 °C overnight at 120 rpm. After enrichment, the mixture was centrifuged at 7000 × g for 15 min and the supernatant was filtered using a 0.22 µm pore-size membrane filter (HIMEDIA) to obtain bacteriophage stock. The filtrates were kept in Ringer Solution (2.25 g NaCl, 0.105 g KCl, 0.12 g CaCl 2 , 0.05 g NaHCO 3 in 500 mL distilled water) (Oxoid™) with a 1:1 ration (v/v) at 4 °C as a working solution for further analysis 37-39 . Bacteriophage titer determination. Titer were determined using the agar overlay assay method 37  www.nature.com/scientificreports/ Bacteriophage application on food samples. Cooked rice, pasteurized milk, chicken meat, and fresh lettuce were used as food samples. Raw chicken meat was cut into pieces (1 cm × 1 cm). Cooked rice and raw chicken meat were placed in 50 mL of Falcon tubes (Corning®) for approximately 1 g for each tube, whereas pasteurized milk (1 mL) was placed in 15 mL of Falcon tubes (Corning®). These samples were sterilized by autoclaving for 15 min at 121 °C to kill all natural bacteria 33 . Meanwhile, fresh lettuces were rinsed with clean water, followed by swabbed with 96% of alcohol on its surfaces. The lettuces were cut into pieces (1 cm × 1 cm) and placed into 50 mL Falcon tubes (Corning®). Then the tubes were exposed to UV light from laminar airflow (ESCO) for about 45 min 44 .
After sterilization, each sample of cooked rice and pasteurized milk were inoculated with 100 μL of mid-log phase bacterial host strain suspensions (B. cereus and B. subtilis) and 100 μL of isolated bacteriophages (BC-S1 and BS-S2) which were diluted to contain MOI of 0.1. While each sample of chicken meat and fresh lettuce were inoculated with 100 μL of mid-log phase bacterial host strain suspensions (ETEC and EHEC) and 100 μL of isolated bacteriophages (ETEC-S3 and EHEC-S4) which were diluted to contain MOI 0.001 for ETEC and MOI 1 for EHEC. All samples were then incubated at 4 °C and 28 °C overnight 45 .
After the incubation, 10 mL of SM buffer was added to each sample and the tubes were vortexed for around 3 min. Each sample was then serially diluted and spread onto LB agar plate, incubated at 37 °C overnight. For positive control, food samples were inoculated with host strain only. For negative control, food samples were inoculated with isolated bacteriophage lysate only. Colonies were counted between 30 and 300 colonies which expressed as colony forming unit per milliliter (CFU/mL) 46 .
Statistical analysis. The data were collected after 3 times of replication and statistical analysis was done using one-way ANOVA followed by Tukey's-B test (SPSS Inc. IBM corporation). The level of difference was defined at P ≤ 0.05, and different letters in each column indicated significant differences from other samples. For control-treatment pairing of each sample, its significant reduction was determined using the paired-samples T-Test with the level of differences defined at P ≤ 0.05 47 .

Conclusions
Four lytic bacteriophages, BC-S1, BS-S2, ETEC-S3, and EHEC-S4 were successfully recovered from soil samples. They were considered highly specific phages with a narrow spectrum of host range, where phage ETEC-S3 was found inefficient against B. cereus, and phage EHEC-S4 had low efficiency against EPEC. This narrow property is desirable because it usually has great specificity to the host. By using TEM, phage BC-S1, and BS-S2 could be categorized as one of the Caudovirales members. These phages showed a significant reduction in food samples on miMOI of 0.1 for both phage BC-S1 and BS-S2, miMOI 0.001 for phage ETEC-S3, and miMOI 1 for phage EHEC-S4 at 4 °C and 28 °C storage temperature. These results showed the potential efficacy of bacteriophage in reducing targeted food spoilage and foodborne bacteria. Thus, it is also promising to be studied further.
Limitation. This study only screened some of the food spoilage and foodborne pathogenic bacteria, also the food that has been assayed is limited, therefore other microbes and food samples need to be explored. On the other hand, it is also should be characterized further for their genomic properties like virulence factor and antibiotic resistance genes also the survival of this bacteriophage in various food processing conditions.

Data availability
The data of this study is available with the corresponding author upon request.

Funding
This study was funded by the Ministry of Education of Indonesia-2021. The funder has no contribution in this study.