Introduction

Before 1980, Cronobacter sakazakii (C. sakazakii) was known as Enterobacter cloacae. Then, till 2007, Enterobacter cloacae was named as Enterobacter sakazakii. It was then reclassified into genus Cronobacter on the basis of its nucleotide sequence (Farmer et al., 1980; Iversen et al., 2007). The Cronobacter genus includes the following seven species: C. sakazakii, C. malonaticus, C. turicensis, C. universalis, C. muytjensii, C. dublinensis, and C. condimenti (Brady et al., 2013; Joseph et al., 2012). C. sakazakii is ubiquitous in nature and thus can be isolated from the environment and food materials and products, including dried foods and water (Healy et al., 2010; Kandhai et al., 2004). C. sakazakii has higher tolerance to desiccation, osmotic stress, and heat stress than other members of Enterobacteriaceae (Asakura et al., 2007), contributing to its survival in desiccation and osmotic stress environments, typical of powdered infant formula.

In recent years, C. sakazakii, as an emerging food-borne pathogen, has gained more and more attention. Genus Cronobacter can cause severe diseases, including bacteremia, sepsis, brain abscess, meningitis, and necrotizing enterocolitis in immunocompromised neonates, especially in pre-term and low birth weight infants (Drudy et al., 2006; Lai, 2001; Nazarowec-White and Farber, 1997; Yan et al., 2012). Besides, genus Cronobacter causes urosepsis, pneumonia, and bacteremia in immunocompromised adults, especially in the elderly population (Hawkins et al., 1991; Lai, 2001; See et al., 2007). In the United States, incidences of 1 Cronobacter infection per 100,000 infants, 8.7 Cronobacter infections per 100,000 low birth weight neonates (Himelright et al., 2001), and 1 Cronobacter infection per 10,660 very-low birth weight neonates (Stoll et al., 2004) have been reported. Although the incidence rate of C. sakazakii infection is low, fatality because of its infection is as high as 80% (Friedemann, 2009). One of the most severe outbreaks of C. sakazakii infection was in a neonatal intensive care unit of France in 1994, which lasted for more than 3 months, infecting 17 neonates and claiming three lives (Caubilla-Barron et al., 2007). Further, 18 cases of (meningitis or) bacteraemia in infants aged 1–11 months have been reported in the United Kingdom by the Food and Agriculture Organization/World Health Organization (FAO/WHO) during 1997–2007, and in 2008, 27 clinical Cronobacter isolates from young children aged 1–4 years have been reported in England and Wales (FAO/WHO, 2008). Besides, the C. sakazakii outbreaks were also reported in Belgium during 1997–1998, in Austria during 2009 –2016, and in France during 2010–2016 (Lepuschitz et al., 2009). Prevalence of infection, high mortality rates, and associated chronic neurological and developmental disorders in many survivors highlight the damaging effects of this organism on infant health (Forsythe, 2005; Lai, 2001). Thus, the International Commission for Microbiological Specifications for Foods (ICMSF), which was formed in 1962 through the action of the International Committee on Food Microbiology and Hygiene as a committee of the International Union of Microbiological Societies (IUMS) and linked to the International Union of Biological Societies (IUBS) and the World Health Organization (WHO) of the United Nations, has ranked C. sakazakii as a ‘severe hazard for restricted populations, life threatening or substantial chronic sequelae of long duration’, and has classified it with Clostridium botulinum, Cryptosporidium parvum, and Listeria monocytogenes (Abbasifar et al., 2014).

Therefore, it is important to minimize the risk of C. sakazakii contamination in foods by developing novel alternative biocontrol agents. Bacteriophages are recognized as safe, host-specific, and effective alternatives for the prevention and/or eradication of food-borne pathogens in foods and their processing environments. In fact, bacteriophages have been applied in the decontamination of livestock; sanitation of contact surfaces and equipment; and biocontrol of raw meats, fresh foods, and vegetables (Endersen et al., 2014; Goodridge and Bisha, 2014), cheese (Carlton et al., 2005), ready-to-eat foods (Bigot et al., 2011), skim milk (Ellis et al., 1973; Endersen et al., 2013), and reconstituted infant formula (Kim et al., 2007). Selected C. sakazakii phages were used to inhibit growth of C. sakazakii in the formula and show high efficiency (Kim et al., 2007). Furthermore, 67 newly isolated C. sakazakii phages have been tested, some of which have reduced C. sakazakii up to 4 log (CFU/mL) in pure broth culture, which shows that the C. sakazakii phages have a great potential of being a biocontrol agent for controlling C. sakazakii in foods (Zuber et al., 2008). In addition, C. sakazakii phages also exhibit excellent efficiency in alleviating Cronobacter-induced urinary tract infections in mice (Tóthová et al., 2011).

These studies demonstrate that bacteriophages are promising natural agents for the control of C. sakazakii. Although a total of 21 genomes of phages infecting C. sakazakii have been published in the National Center for Biotechnology Information (NCBI) database (https://www.ncbi.nlm.nih.gov/genome) so far, the library of C. sakazakii phages and their genomic information are still limited for developing biocontrol agents, because C. sakazakii strains are of high diversity and tend to become phage-resistant. Therefore, new broad host range phages must be isolated, and phage cocktails must be created as broad-spectrum food biocontrol agents against C. sakazakii.

In this study, human stool samples and sewage samples were used to isolate broad-spectrum virulent phages against C. sakazakii. In all, 10 bacteriophages infecting C. sakazakii were isolated and purified, and their morphological features, dynamics of infection, and host range were characterized. In addition, genomes of two broad-spectrum phages (EspYZU05 and EspYZU08) were completely sequenced to understand their characteristics. On the basis of these results, cocktail of virulent phages were prepared and evaluated for their potential and efficacy in the biocontrol of C. sakazakii strains in the liquid broth and milk medium under different temperatures and pH values to evaluate their possibility of being novel and efficient biocontrol agents.

Materials and Methods

All the chemicals used in this study, except for nutrient broth, were of analytical grade and purchased from Sangon Biotech Co. Ltd. (Shanghai).

Bacterial strains and cultivation

C. sakazakii strains CICC 21560, CICC 21545, CICC 21569, CICC 21673, and CICC 22919 were used as bacterial hosts for isolating phages in this study. These strains were purchased from the China Industrial Culture Collection Center. The strains were stored at -80°C and routinely based on nutrient agar and nutrient broth (Hangzhou Microbial Reagent Co. Ltd.) at 37°C.

Collection of samples

A total of 100 samples of human stool were collected from patients suspected of C. sakazakii infection in the Affiliated Hospital of Yangzhou University, Yangzhou, Jiangsu, China after signing the consent. Apart from these, 12 raw sewage samples were collected from Kangyuan Dairy Co. Ltd. (Yangzhou, Jiangsu, China).

Isolation and purification of phages

In order to isolate bacteriophages, we homogenized 25 mL of fresh sewage or 5 g of stool sample with 45 mL of sodium chloride–magnesium sulfate (SM) buffer (NaCl, 5.8 g; MgSO₄•7H₂O, 2 g; Tris-HCl 1 mol/L [pH 7.5], 50 mL; gelatin, 0.1 g: dissolved in distilled water to a final volume of 1 L). After overnight incubation at 4°C, the homogenized sample was centrifuged at 5,000 × g for 10 min at 4°C, and the supernatant was filtered using sterile 0.22-μm membrane. Afterward, 5 mL of filtrate and 100 μL of host bacteria (C. sakazakii) suspension in logarithmic growth phase were aseptically added to a tube with 5-mL nutrient broth incubated overnight at 37°C with continuous shaking at a speed of 120 rpm/min. After incubation, the broth was centrifuged at 5,000 × g for 10 min at 4°C. The phage suspended in the semen was filtered through a sterile 0.22-μm membrane to remove residual bacterial cells.

Thereafter, the presence of viable infective phages in supernatant was tested by a two-layer plating method. The supernatant (100 μL) was mixed with host bacteria (100 μL) in early logarithmic growth phase and added to a tube with 5 mL of soft nutrient agar (0.7% agar), tempered to 37°C, and thereafter poured onto the surface of a nutrient agar plate (2% agar). The plates were incubated overnight at 37°C and examined for phage plaques. Extract a single phage plaque and resuspend it in SM buffer. The isolation processes were repeated for three to five times to purify individual phages.

Amplification and collection of phages

Inoculation of 5 mL of the mixture of host bacteria and bacteriophages into 500 mL of fresh nutrient broth was performed and shaken overnight at 37°C at a speed of 120 rpm/min. After phage propagation and amplification, the lysates were treated with DNase I and RNase A with a final concentration of 1 μg/mL at 37°C for 30 min and then with 1 mol/L NaCl in ice for 1 h. The host cell fragments were removed at a 10,000 × g centrifugation of 10 min at 4°C. Phage particles were concentrated in 10% (w/v) polyethylene glycol 8,000 by overnight incubation in ice. The phages were pelleted by 11,000 × g centrifugation for 15 min at 4°C and resuspended in 1 mL of SM buffer. The phages were treated with 1 mL of chloroform and centrifuged at 3,000 × g for 15 min at 4°C. Phages dispersed in the upper aqueous phase were collected. The aqueous phase was overlaid on a CsCl step gradient (density = 1.4, 1.5, and 1.7 g/mL, 1 mL each step) in 5-mL centrifuge tube and horizontal centrifuge at 22,000 × g for 2 h at 4°C. The phage band (density between 1.5 and 1.7 g/mL) was drawn through the wall of centrifuge tube by using a syringe and stored at 4°C for further experiments.

Determination of phage titer

First, the purified phages were diluted for 10 times with SM buffer. Thereafter, 100 μL of phage suspension was mixed with 100 μL of host bacterium suspension. The mixture was added to 5 mL of nutrient broth incubated at 37°C for 12 h with continuous shaking at 120 rpm/min. The culture medium was centrifuged at 5,000 × g for 10 min at 4°C, and the supernatant was filtered through a sterile 0.22-μm membrane filter. Subsequently, 100 μL of filtrate (phage suspension) was serially diluted with a gradient to 10-fold, mixed with 100 μL of host bacterium suspension (~10⁹ CFU/mL), and incubated for 10 min at 37°C. Then the mixture was added to 5 mL of soft agar and spread on nutrient broth agar plates. The titers of the phages were determined by the soft agar overlay method. All assays were carried out in triplicate.

Transmission electron microscope

Phage morphology was observed by TEM (Phenom XL G2, the Netherlands) analysis. CsCl phage suspension was dropped on the membrane side of 400 mesh copper grid. After 10 min of adsorption, excess solution was removed with a bibulous paper. The copper grid was treated for 2 min in a drop of 2% phosphotungstic acid, and excess liquid was removed by infrared light. Finally, the specimens were examined and the morphology and dimensions of phages were recorded with a Tecnai-12 TEM.

Host range activity

Soft agar, 5 mL, containing 100 μL of host bacteria was overlaid on 1.5% nutrient agar plates. Then 10 μL of phage suspension (~10⁹ PFU/mL) was spotted on the overlaid plates and dried with blotting paper. Petri dish was cultured for 8–12 h at 37°C. The lysis activity of the test bacteria by phages was determined by counting the plaques of spots.

Effect of heat treatment on phage infectivity

In order to determine the effect of temperature on phages, we placed 500 μL of phage suspension in water bath at 50, 60, and 70°C. Subsequently, 100 μL of phage suspension was immediately obtained for phage titer measurement for 20, 40, and 60 min as described in Section 2.5. Measurement was replicated for three times, and the average values were used for analysis.

Effect of pH on phage infectivity

Effect of pH on phage activity was determined in nutrient solution with different pH values (pH 3–11). In general, 50 μL of phage suspension was mixed with 950 μL of nutrient broth adjusted to pH of 3–11 and incubated at 37°C for 2 h. Subsequently, as described in Section 2.5, 100 μL of phage suspension was obtained immediately for phage titer measurement. All assays were performed in triplicate.

Optimization of MOI

MOI is a ratio of virus particles to potential host cells. The host cells were infected with phages in different ratios (0.001, 0.01, 0.1, 1, and 10 PFU/CFU) and incubated at 37°C for 6 h. The culture medium was centrifuged at 5,000 × g at 4°C for 10 min. The supernatant was filtered through sterile 0.22-μm membrane filter, and phage titer was measured as described in Section 2.5. The MOI resulting in the highest phage titer within 6 h was regarded as optimal MOI.

One-step growth curve of phage

First, 100 μL of phages (~10⁸ PFU/mL) and their host bacteria (~10⁷ CFU/mL) were mixed and allowed to adsorb for 10 min at 37°C. Here, the C. sakazakii strain of CICC 21560 was used as host for EspYZU01 and EspYZU02. The C. sakazakii strain of CICC 22919 was used as host for EspYZU03, EspYZU04, EspYZU05, EspYZU06, EspYZU07, EspYZU08, EspYZU09 and EspYZU10. The culture was centrifuged at 5,000 × g for 30 s at 4°C. The pelleted cells were washed twice with pre-warmed nutrient broth, resuspended in 5 mL of nutrient broth, and incubated at 37°C. The bacteriophage titer was determined by double agar plate method. The samples were taken every 10 min in 0–2 h and every 15 min in 2–4 hours. A one-step growth curve was drawn with phage processing time as the abscissa and phage titer as the ordinate. During the incubation period, the burst size was calculated by the ratio of the final count of released phage particles to the initial count of infected bacterial cells.

Genome sequencing and analysis

DNA extraction and purification

The DNA extraction steps are the same as provided in the instructions of Ezup Spin Column Super Plant Genomic DNA Extraction Kit (Sangon Biotech, Shanghai, China). Briefly, first, 500 μL of purified phages was mixed with 20 μL of 500 mmol/L ethylenediaminetetra acetic acid, 30 μL of 10% SDS, and 3 μL of 10 mg/mL proteinase K and incubated at 56°C for 1 h. Isovolumetric chloroform:isoamyl alcohol:phenol (25:24:1) mixture was added, mixed thoroughly, and centrifuged at 12,000 × g for 5 min at 4°C. The upper layer was carefully transferred to a new sterile tube. This step was repeated using equal volumes of chloroform:isoamyl alcohol (24:1) mixture, which was centrifuged at 12,000 × g for 5 min at 4°C. After incubation at –20°C for 1 h, DNA was precipitated from the solution with isopropanol. The solution was centrifuged at 4°C at 12,000 × g for 10 min, and the supernatant was separated. The precipitate was washed twice with 70% ethanol, dried at room temperature, resuspended in 20-μL TE buffer (pH 7.4), and dissolved at 37°C for 30 min. The quality and quantity of DNA was evaluated by GeneQuant (Eppendorf, Germany) and by running DNA on an agarose gel by electrophoresis followed by visualization.

Genome sequencing and annotation

The genomic DNA of phage was sequenced by Novogene Biology Information Technology Co. Ltd. (Beijing, China). A polymerase chain reaction (PCR)-free sequencing library of inserts of approximately 500 bp was constructed, and the Illumina HiSeq^TM 2000 sequencing platform was used for sequencing. The obtained raw sequencing data were filtered, and sequences containing the linker, primer, and low-quality data were removed. The clean data were used for subsequent analysis. The phage genome sequence was assembled using the SOAPdenovo software, and assembly conditions were adjusted to obtain the best stitching results. The amino acid sequences of the target species were compared with non-redundant (NR) database, and blast technology was used to locate the target species’ genes and their corresponding functional annotation information (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Given that each sequence may have more than one alignment, an optimal alignment result was retained as an annotation for the gene in order to ensure its biological significance. Homology alignment of the target and reference genomes was performed using the MUMmer software. A genome-wide map of phage was constructed using CGView (http://wishart.biology.ualberta.ca/cgview/).

Application of phages

Preparation of phage cocktail

The method for preparing phage cocktail is as follows: 1-mL phage suspension with a titer of ~10⁸ PFU/mL is mixed, and then stored at 4°C until take into use. Phage cocktail-1 contains EspYZU01 and EspYZU05; cocktail-2 contains EspYZU02, EspYZU03, and EspYZU07; and cocktail-3 contains EspYZU01, EspYZU03, EspYZU08, EspYZU09, and EspYZU10. The C. sakazakii cocktail was prepared by mixing 1 mL of each host bacteria at a concentration of ~10⁷ PFU/mL, and used immediately.

Application of phage against C. sakazakii in nutrient broth

First, 100 μL of phage cocktail or EspYZU08 (~1×10⁸ PFU/mL) was mixed with 100 μL of C. sakazakii cocktail (~1×10⁷ CFU /mL). The mixture was then inoculated into 4.8 mL of nutrient broth and cultured at 37°C, 25°C, and 4°C with continuous shaking at a speed of 120 rpm/min. Optical density values at 600 nm (OD_{600 nm}) were measured every 3 h at 25°C and 37°C and every 24 h at 4°C. Nutrient broth containing 100 μL of SM buffer instead of bacteriophage served as a positive control. All analyses were performed in triplicate.

Application of phage against C. sakazakii in milk

Phage cocktail or EspYZU08 (~1×10⁸ PFU/mL), 5 mL, was mixed with 5 mL of C. sakazakii cocktail (~1×10⁷ CFU/mL), which was mixed by five strains of CICC 21560, CICC 21545, CICC 21569, CICC 21673, and CICC 22919 in the same ratio. The mixture was then inoculated into 40 mL of milk and incubated at 37°C, 25°C, and 4°C with continuous shaking at a speed of 120 rpm/min. Standard plate count was performed on nutrient agar every 3 h at 25°C and 37°C and every 24 h at 4°C to quantify surviving cells. Milk containing 5 mL of SM buffer instead of bacteriophage was used as a positive control. All experiments were carried out in triplicate, and the bacterial concentration was expressed as mean CFU/mL count and standard deviation.

Results

Isolation and purification of phages

A total of 10 phages infecting C.sakazakii strains were isolated from human stool and sewage samples and were marked as EspYZU01–EspYZU10. All 10 phages formed visible and uniform size plaques on the host strain (Figure 1A). Among these 10 phages, EspYZU01, EspYZU02, EspYZU06, EspYZU07, EspYZU08, EspYZU09, and EspYZU10 formed clear plaques of 0.5–1 mm in diameter. EspYZU03 and EspYZU04 formed blurry plaques of approximately 1 mm in diameter, and EspYZU05 formed a clear plaque of approximately 4 mm in diameter.

Figure 1. Characterization of (A) plaque and (B) microscopic morphological features of Cronobacter sakazakii phages. 01: EspYZU01; 02: EspYZU02; 03: EspYZU03; 04: EspYZU04; 05: EspYZU05; 06: EspYZU06; 07: EspYZU07; 08: EspYZU08; 09: EspYZU09; 10: EspYZU10. (A) Images of phage plaques are magnified four times and shown with red squares. (B) Morphology and dimensions of phages were examined with a Tecnai-12 transmission electron microscope.

The phages were then purified and collected by discontinuous Cesium chloride (CsCl) density gradient centrifugation. A visible bacteriophage band of 1.45–1.50 g/mL CsCl was obtained with a titer of 10¹⁰ –10¹¹ PFU/mL.

Morphology of phages

Phage morphology was characterized by transmission electron microscopy (TEM). The results showed that 10 Cronobacter phages exhibited four types of morphology (Figure 1B). EspYZU01 and EspYZU09 had an elongated polyhedron head with a diameter of ~70 nm × 120 nm (L/W = 1.7) and a tail length of ~120 nm, indicating that the phages belonged to the Myoviridae family. EspYZU02 and EspYZU08 had an isometric polyhedron head with a diameter of ~90 nm, tail length of ~120 nm, and a contractile tail sheath, indicating that the phages also belonged to the Myoviridae family. EspYZU03, EspYZU04, and EspYZU07 had an isometric polyhedron head but without tail sheath, indicating that the phages also belonged to the Myoviridae family. EspYZU05 had an isometric polyhedron head with a diameter of ~55 nm and a tail length of ~18 nm, indicating that the phages belonged to the Podoviridae family. EspYZU06 had an isometric polyhedron head with a diameter of ~60 nm but without a tail, indicating that the phages belonged to the Tectivirus family. EspYZU10 had an isometric polyhedron head with a diameter of ~130 nm and a tail length of ~250 nm but without tail sheath, indicating that the phages belonged to the Stylovinidae family.

Host range of phage

The results of host range of phages are listed in Table 1. EspYZU02 had the highest specificity and could only infect the CICC 21560 strain. EspYZU08 had the broadest infection spectrum and could infect all five C.sakazakii strains. The eight other phages could infect two to four C.sakazakii strains (Table 1).

Characterization of phages

Effect of temperature and pH on phage infectivity

The effect of temperature on phage infectivity was tested by exposing phages to a range of different extreme temperatures. When the phages were exposed to 50°C for 1 h, infectivity retained by all phages was quite well. At 60°C for 1 h, the infectivity of EspYZU01, EspYZU07, EspYZU08, and EspYZU10 declined slightly (<20%); that of EspYZU02, EspYZU03, EspYZU04, and EspYZU6 declined by 20–44.8%; and that of EspYZU05 and EspYZU9 declined by >50%. At 70°C, phage infectivity declined sharply. EspYZU01, EspYZU02, EspYZU04, and EspYZU5 lost their infectivity after 40 min; EspYZU03 and EspYZU9 lost their infectivity after 20 min; and only EspYZU06, EspYZU07, EspYZU08, and EspYZU10 retained 22.1–41.7% of their infectivity after 1 h (Figure 2A). The results indicated that EspYZU07, EspYZU08, and EspYZU10 have the best thermal stability.

Figure 2. Effects of (A) temperature and (B) pH on phage infectivity. 01: EspYZU01; 02: EspYZU02; 03: EspYZU03; 04: EspYZU04; 05: EspYZU05; 06: EspYZU06; 07: EspYZU07; 08: EspYZU08; 09: EspYZU09; 10: EspYZU10; (A) The effect of temperature on the viability of phage in nutrient broth at 50–70°C for 20–60 min. (●), (■), and (▲), respectively, represent the infectivity of phage exposed at 50°C for 20–60 min, 60°C for 20–60 min, and 70°C for 20–60 min. (B) The effect of pH on the viability of phage in nutrient broth at pH 3–11 and 37°C for 2 h. Each assay was conducted in triplicate, and the values were expressed as mean ± standard deviation.

The effect of pH on phage infectivity was tested by exposing phages to pH ranging from 3.0 to 11 for 2 h at 37°C. The infectivity retained by all phages was quite well if they were exposed to an environment having pH 5.0–9.0 but declined sharply at pH < 5.0. At pH < 4 or >10, EspYZU01, EspYZU06, EspYZU07, EspYZU08, and EspYZU10 retained their high infectivity. Good infectivity of EspYZU03, EspYZU04, EspYZU05, and EspYZU09 was retained in alkaline environment (pH = 7–10) but declined sharply at pH < 4 (Figure 2B). The results indicated that EspYZU01, EspYZU06, EspYZU07, EspYZU08, and EspYZU10 had the best pH stability.

Optimal multiplicity of infection (MOI) of phages

The MOI of all phages were determined as described in Section 2.10, and the results are listed in Table 2. EspYZU02, EspYZU03, and EspYZU04 had the maximal MOI of 10, and EspYZU06 and EspYZU10 showed the minimum MOI of 0.001.

One-step growth curves of phages were obtained by propagation on C.sakazakii at 37°C (Figure 3). The latent periods of phages occurred from 10 to 45 min. The rising period began with average burst size of 65–439 phage particles per bacterium (Table 3). Among the phages, EspYZU09 had the shortest latent period of 10 min, and EspYZU06 showed the longest latent period of 45 min. EspYZU10 had the maximal burst size of 439, and EspYZU08 demonstrated the second maximal burst size of 366.

Figure 3. One-step growth curve of phages. 01: EspYZU01; 02: EspYZU02; 03: EspYZU03; 04: EspYZU04; 05: EspYZU05; 06: EspYZU06; 07: EspYZU07; 08: EspYZU08; 09: EspYZU09; 10: EspYZU10. Each assay was conducted in triplicate, and the values were expressed as mean ± standard deviation.

Genome analysis of phage

Among the phages, EspYZU05 and EspYZU08 infected the same host of CICC 21569 and showed broad host range. However, they presented remarkably different morphologies; hence, we sequenced and analyzed the whole genomes of EspYZU05 and EspYZU08. The genome of EspYZU05 was obtained with a size of 41723 bp, G+C content of 55.69%, gene coding percentage of 93.04%, and open reading frames (ORFs) of 47 (Figure 4A). The genome data of EspYZU05 was deposited in GenBank under the accession number MW882933; annotation of ORFs is shown in Table S1 provided in supplementary material. The genome of EspYZU08 with a size of 145582 bp, G+C content of 46.76%, gene coding percentage of 42.03%, and ORF of 103 was obtained (Figure 4B). The genome data of EspYZU08 was deposited in GenBank under the accession number MW882934; annotation of ORFs is shown in Table S2 provided in supplementary material. Genomes of both EspYZU05 and EspYZU08 were double-stranded DNAs, and no tRNA gene was analyzed. Regarding the similarity of phage proteins with those from the NCBI database, the genome of EspYZU05 was found to be highly similar to that of C. sakazakii phage vB_CskP_GAP227 (KC107834.1), with 96% identity and 98% coverage rate. The genome of EspYZU08 was very similar to Salmonella phage PVP-SE1 (GU070616.1) with 97% identity and 82% coverage rate.

Figure 4. Genome pattern of phages EspYZU05 and EspYZU08. (A) Genome pattern of phage EspYZU05 (accession number is “MW882933”); (B) Genome pattern of phage EspYZU08 (accession number is “MW882934”). Circles display (from the outside): (1) ORFs transcribed in clockwise or counterclockwise direction; (2) G+C% content. Values >42.97% (average) are outward peaks, and those <42.97% are inward peaks; (3) GC skew (G−C/G+C, in a 1-kb window and 0.1-kb incremental shift). Values >0 are inward peaks, and those <0 are outward peaks; (4) Physical map is scaled in kbp (for interpretation of references to color in this figure, the reader is referred to the web version of this article).

The functional ORFs of phage genomes could be classified into structure (e.g., membrane protein, scaffolding protein, capsid/head protein, head–tail connector protein, tail tubular protein, minor tail protein, tail sheath protein, and tail attachment protein), packaging (e.g., ribonucleotide reductase), DNA manipulation (e.g., DNA helicases, DNA polymerase, DNA ligase, DNA primase, DNA maturase, and DNA exonuclease), transcription (e.g., RNA polymerase and RNA ligase), and additional functions (e.g., nicotinamide–nucleotide adenylyltransferase and transposase protein). Furthermore, some host lysis-related proteins were found, such as lytic glycosylase (ORF39) and endolysin (ORF45) in EspYZU05 and colanic acid degrading protein (ORF18) in EspYZU08, which contribute to infecting and lysing the host cell. However, many products of predicted ORFs in genomes remain hypothetical proteins, these may result from the insufficient annotation data of C. sakazakii bacteriophage genomes.

Furthermore, the safety of phages was assessed on the basis of genomes, and no gene for toxins and antibiotic resistance was detected.

Bacterial challenge test of phage and cocktail in liquid broth

In order to evaluate the best antibacterial efficacy of phages, we measured the infectivity of EspYZU08 and phage cocktails in liquid broth at different temperatures (Figure 5). Cocktail-3 (EspYZU01 + EspYZU03 + EspYZU08 + EspYZU09 + EspYZU10) could inhibit the growth of C.sakazakii for 12 h at 37°C, 24 h at 25°C, and 12 days at 4°C. The inhibition ratio of I_{12 h}, I_{24 h}, and I_{12 d} reached 99.20%, 98.69%, and 72.40%, respectively. Compared with other phages and cocktails, cocktail-3 presented the best antibacterial effect. Thus, cocktail-3 was used to further evaluate its antibacterial effect in food.

Figure 5. Growth inhibition of C. sakazakii using phage EspYZU08 or phage cocktail in nutrient broth at different temperatures. Cocktail-1: EspYZU01 + EspYZU05; cocktail2: EspYZU02 + EspYZU03 + EspYZU07; cocktail-3: EspYZU01 + EspYZU03 + EspYZU08 + EspYZU09 + EspYZU10. (■) represents C. sakazakii grown in the absence of phage; (●) represents C. sakazakii grown in the presence of phage EspYZU08 or phage cocktail (~1 × 10⁸PFU/mL). Growth inhibition of C. sakazakii was determined by CFU/mL counts. Each assay was conducted in triplicate, and the values were expressed as mean ± standard deviation.

Application of phage cocktail in milk

In order to verify the potential of phages as a novel biocontrol agent against C.sakazakii in food, we used cocktail-3 (EspYZU01 + EspYZU03 + EspYZU08 + EspYZU09 + EspYZU10) with a titer of ~1×10⁸ PFU/mL to evaluate antibacterial effect in milk at different temperatures (Figure 6). After cocktail-3 was added, the total viable counts of C.sakazakii in milk decreased to the minimum value for 6 h at 37°C, 9 h at 25°C, and 5 days at 4°C and reduced by 2.92 lg CFU/mL, 2.97 lg CFU/mL, and 1.64 lg CFU/mL compared with the initial values and by 4.57 lg CFU/mL, 4.25 lg CFU/mL, and 3.43 lg CFU/mL compared with the control (no phage), respectively. In addition, addition of cocktail-3 significantly decreased the population of C.sakazakii (P < 0.001) relative to the control within a certain time frame. This result suggested that cocktail-3 could remarkably inhibit growth of C.sakazakii in milk for 6 h at 37°C–9 h at 25°C and 5 days at 4°C.

Figure 6. Growth inhibition of C. sakazakii using phage cocktail-3 in milk. Cocktail-3 was mixed with EspYZU01, EspYZU03, EspYZU08, EspYZU09, and EspYZU10. (■) represents C. sakazakii grown in the absence of phage; (●) represents C. sakazakii grown in the presence of phage cocktail (~1 × 10⁸ PFU/mL). Growth inhibition of C. sakazakii was determined by CFU/mL counts. Assays were performed in triplicate and bacterial concentrations were expressed as mean ± standard deviation.

Discussions

C. sakazakii, generally found in contaminated infant milk formula powders, is a fatal food-borne pathogen with high mortality rates (Drudy et al., 2006). Bacteriophage is considered as an alternative and promising approach to control C. sakazakii in foods. In this study, we isolated and characterized 10 Cronobacter phages with host lysis activity from sewage and stool samples of patients suspected of C. sakazakii infection. Sewage is a primary niche for Enterobacteriaceae; thus, the isolation of Cronobacter phages from effluent environments is not uncommon (Kim et al., 2007; Zuber et al., 2008). After morphological observation through TEM, 10 phages were considered to belong to Myoviridae, Podoviridae, Tectivirus, and Siphoviridae families. The phage susceptibility was assessed using five C. sakazakii strains. EspYZU08 had the broadest host range because it infected all five strains. The broad host range capabilities of five combined C. sakazakii phages show an infection profile extending across several genera (Zuber et al., 2008). Thus, the C. sakazakii phage of EspYZU08 may have good application prospect to control multiple pathogens in food. In addition, given that phages EspYZU08 and EspYZU05 have lytic lifestyles and do not possess genes for toxic proteins, they meet the required properties of phages intended for biocontrol applications. Although a transpose was found in phage EspYZU08, the transpose was proved to be safe for human cells and even could be used for clinical trial (Magnani et al., 2018; Zhang et al., 2021).

Environmental stability is essential for phages to be used as a biocontrol in foods. The common environmental pressure includes low pH and high temperature. The environmental stability of 10 phages was tested. It was found that all phages retained their maximum infectivity after exposure to pH ranging from 5 to 9, but their infectivity declined sharply at pH < 5. The pH stability of these phages was similar to some other C. sakazakii phages. C. sakazakii phages leB, leE, and leN retain their activity after exposure to pH 6–10, but no viable phages could be observed when exposure to both pH = 2 and 4 for 1 h (Endersen et al., 2017). The C. sakazakii phage PBES 02 also retained its infectivity after exposure to pH = 6–10, but its infectivity decreased at pH < 5 (Lee et al., 2016b). However, some phages have very high pH stability, such as coliphage λ, which shows no remarkable decrease in titer at pH = 3–11 (Jepson and March, 2004).

Temperature also plays a fundamental role in the survival of phages. High optimal temperatures are thought to extend the latent period, whereas low optimal temperatures are often thought to result in reduced multiplication rate (Tey et al., 2009). In this study, phages retained their infectivity at 50°C and showed slightly decreased infectivity at 60°C following a 1-h challenge. However, at 70°C, many phages were inactivated following incubation for 1 h. The thermostability of these phages was also similar to that of some other C. sakazakii phages. The C. sakazakii phages leB, leE, and leN retained their infectivity between 4°C and 50°C, and no viable phages could be recovered from the lysates exposed to 60°C, 72°C, or 90°C for 1 h (Endersen et al., 2017). The infectivity of C. sakazakii phage PBES 02 is retained after exposure to 4–55°C for 1 h but decreased sharply (75% lost) at >65°C for 1 h (Lee et al., 2016b).

Clarifying the genomic information of each bacteriophage is essential to ensure the specificity and safety (without virulence factors) of bacteriophage as a biological therapeutic agent (Brüssow et al., 2004; Faruque and Mekalanos, 2012). Besides, genome sequencing helps to further understand the phage–host interactions and provide necessary information to further exploit their biological therapeutic properties. Thus, the genomes of EspYZU05 and EspYZU08 were sequenced and analyzed, and no gene for toxins and antibiotic resistance was detected. However, some endolysin-supporting proteins were found, such as lytic glycosylase, endolysin, and colanic acid-degrading protein, which support the infection and lysis of host cell.

In this study, the lysis activity of these phages for food application was demonstrated using phage cocktail (~1×10⁸ PFU/mL) against C. sakazakii (~1×10⁷ CFU/mL) in milk. The results showed that cocktail-3 (EspYZU01 + EspYZU03 + EspYZU08 + EspYZU09 + EspYZU10) presented the best efficacy and controlled C.sakazakii for 6 h at 37°C, 9 h at 25°C, and 5 days at 4°C. Relative to initial values, the population of C.sakazakii was reduced by 2.92 lg CFU/mL (44.92%) at 37°C, 2.97 lg CFU/mL (45.69%) at 25°C, and 1.64 lg CFU/mL (25.23%) at 4°C. The Cronobacter phages leB, leE, and leN were combined as a part of phage cocktail (~3×10⁸ PFU/mL) to assess their ability to inhibit the growth of C. sakazakii (~1×10⁴ CFU/mL) in four different brands of infant formula. The C. sakazakii concentrations were reduced to below the detection limit (<10 CFU/mL) in 5-h incubation when challenged with phage cocktail, and this level of inactivation was maintained over the 20-h challenge (Endersen et al., 2017). The Cronobacter phage CR5 at an MOI of 10⁵ was added to a sample containing C. sakazakii, and the bacterial strain was lysed at 2 h and never recovered up to 10 h (Lee et al., 2016a). When the Cronobacter phage PBES 02 with an MOI of 10⁵ was added to infant formula containing C. sakazakii, the bacteria were completely eliminated in 6 h (Lee et al., 2016b). The Cronobacter phage Dev2 completely killed the bacteria at a high initial MOI (10² CFU/mL bacteria and 10⁸ PFU/mL phages) in LB medium and reconstituted milk formula; similar results were observed at 20°C, 30°C, and 37°C (Kajsík et al., 2014).

In general, we used a low MOI of 10, and the biocontrol results were similar to those of some other C. sakazakii phages. The level of C. sakazakii contamination in powdered infant formula is very low (<1 bacterial cell/100 g) (Holý and Forsythe, 2014). However, the contaminating levels of C. sakazakii used in this study were much higher than those typically found in powdered infant formula, demonstrating the efficacy of this phage cocktail to be explored further.

Conclusions

A total of 10 Cronobacter phages (EspYZU01–EspYZU10) were isolated from sewage and human stool samples. After morphological observation and characterization, the genomes of phages EspYZU05 and EspYZU08 were analyzed, and no toxins and antibiotic resistance genes were detected. The phage cocktail-3 (EspYZU01 + EspYZU03 + EspYZU08 + EspYZU09 + EspYZU10) presented antibacterial efficacy against C.sakazakii in milk for 6 h at 37°C, 9 h at 25°C, and 5 days at 4°C. These results suggest that this phage cocktail may be used to develop a novel phage biocontrol agent against C. sakazakii in dairy and its production environment.