Preparation of model phages and host bacteria
All host strains were provided by ATCC (American Type Culture Collection) and NBRC (NITE Biological Resource Centre). Their characteristics and growing conditions are listed in Table 1. Just before the experiment started, each strain was grown twice consecutively in tryptic soy broth (TSB), activated at 26 °C or 37℃ for 24 hours of incubation. On finishing the incubation period, bacteriophages were spread germ-free on sterile soy agar (TSA) each, and incubated in an aerobic manner at 26 °C or 37°C for 24 hours. Grown colonies were subsequently selected and re-cultured as described previously from a single plate of each strain. Once incubated, the slants were saved at 4 –5 °C and served as a working slant for the entire project.
DNA extraction, amplification and 16S-rRNA sequencing
The host bacteria were grown overnight in a liquid Nutrient Broth at 26°C (P.svringe NBRC14084 ) or 37°C (E.coli ATCC13706 and E.coli ATCC15597 ). The CTAB method was used for the isolation of genomic DNA as described by Akinyemi and Oyelakin.19 The 16S-rRNA gene was amplified using template DNA extracted from genomic DNA. The PCR mix comprised of 5 µL of 10 × buffer, 1 µL of 2.5 mMdNTPs, 2 µL 5 mM forward and reverse primers, 1 µL of 5 units/µL Taq containing 1 µL of template DNA and 40 µL of distilled water to make-up 50 µL reaction mix. Amplification of the nearly complete 16S-rRNA gene with universal 27f (5′-AGAGTTTGCCTGGCTCAG-3′) and 1492r (5′-TACGGCTACCTTGTTACGACTT-3′) primers was carried out by PCR in 30 cycles denaturation (94℃, 60 s), annealing (55℃, 60 s), and elongation (72℃, 30 s), and a final polymerization step (72℃, 10 min). Detected by 2% agarose gel electrophoresis, the PCR products were purified using a Direct PCR purification kit in accordance with the manufacturer's guidelines (B639289, Shanghai, China). The obtained final products were sent to Qingke Biotechnology Co., Ltd (Beijing, China) for sequencing.
Effect of gaseous ozone to viral substitutes in water, gas phase
A sketch of the apparatus used in this research are presented as Fig. 1. An approximately 1.5-liter stainless steel chamber (15 cm × 10 cm × 10 cm) was installed as a sterilization box. On one hand the ozone gas flowmeter (WT-RSL-YT, Nanjing China) was connected with the ozone generator (ARDM0.3 Shanghai China), on the other hand it was connected to the top of the box through a polypropylene-random pipe and extended to bottom. To control for bias caused by air, the sterilization box was equipped with a vacuum air pump on the side.
Stainless steel sheets were autoclaved and put in 5 mL of each virus stock solution. Three model phages finally held the following titer: 7.21 log units, 6.68 log units and 7.51 log units for φ6, φX174 and MS2, respectively. To determine ozone gave the effect on bacteriophage in gas phase, the following procedures were used. The contaminated stainless steel sheets were placed horizontally in the center of the disinfection box. After the multifunctional generator started, the transfer switch was not turned on until 200 mg/h of gaseous ozone produced steadily, avoiding excessive or insufficient amount of gas to increase experiment bias. Finally, the samples were exposed to 200 mg/h gaseous ozone for 0, 1, 2, 3, 4, 5 or 6 min. To assess whether and how resistance of bacteriophages was affected and changed in water phase, we exposed the samples to water filled with ozone and injected stable gaseous ozone for 0, 1, 2, 3, 4, 5 or 6 min. Quantified the effect of gaseous ozone to viral substitutes by log reduction and inactivation following varied time treatment with constant concentrations of gaseous ozone.
Inoculation of samples
A batch of foods and packaging materials (Table 2) were measured and placed in a taker-bag (50 mm × 100 mm). For inoculation, 5 mL of the viral stock solution was added to those bags containing foods and packaging materials. Samples and inoculum were mixed with a slow shaking of the bag for 30 minutes allowing the sample to be uniformly infected while preventing the bag from breaking. Notably, foods were each inoculated superficially, avoiding scarred foods as this is a known way to harbor and internalize pathogens.20 Rather than concentrating on inactivating the bacteria inside the product, we focused only as much on the bacteria on the surface of the product in this study. Samples were loaded with target bacteria at a concentration of 4-6 log units. Once infected, samples were picked up from the bags and counted for bacteria which verified the initial concentration of inoculated pathogens. Subsequently, the infected samples were dried in a cabinet at 22°C for 2 h before ozone gas treatment. The samples were dried, and left at 4°C overnight ensuring the virus particles were dry and adhered.
Treatment of samples with gaseous ozone
The results in previous chapter established theeffect of gaseous ozone to viral substitutes in water, gas phase. In this chapter, therefore, moves on to discuss the effect on samples with ozone at 3,5,6 min for infected samples by φ6, φX174 and MS2, respectively. When a gas flow with 200 mg/h of ozone was passed through the gas flowmeter and detected. The samples were quickly pushed into the box laterally and then disinfected for 3, 5 and 6 min based on our previous tests, respectively. After disinfection, not only did the excessive ozone in the box required released from the bottom, it also needed a complete cleaning and disinfection program. Only in this way can we better avoid bias caused by interactions between samples.
Statistical analysis
The experimental data were transformed to log values, then log reduction values were expressed for each sample group, by using the initial log value of each pathogen minus the processed log value. For inactivation efficacy, the raw data were collated and presented as a decreased percentage of virus titer as accounted by the formula:

Data management and analysis were performed using SPSS v.20 (SPSS, Chicago, IL, USA), Between-mean differences were calculated using Duncan’s multiple range tests. Results with P values of < 0.05 were considered significant and indicated with different letters. Each experiment was repeated five times independently removing the maximum and minimum values.