wksl3, a New Biocontrol Agent for Salmonella enterica Serovars Enteritidis and Typhimurium in Foods: Characterization, Application, Sequence Analysis, and Oral Acute Toxicity Study

Abstract

Of the Salmonella enterica serovars, S. Enteritidis and S. Typhimurium are responsible for most of the Salmonella outbreaks implicated in the consumption of contaminated foods in the Republic of Korea. Because of the widespread occurrence of antimicrobial-resistant Salmonella in foods and food processing environments, bacteriophages have recently surfaced as an alternative biocontrol tool. In this study, we isolated a virulent bacteriophage (wksl3) that could specifically infect S. Enteritidis, S. Typhimurium, and several additional serovars. Transmission electron microscopy revealed that phage wksl3 belongs to the family Siphoviridae. Complete genome sequence analysis and bioinformatic analysis revealed that the DNA of phage wksl3 is composed of 42,766 bp with 64 open reading frames. Since it does not encode any phage lysogeny factors, toxins, pathogen-related genes, or food-borne allergens, phage wksl3 may be considered a virulent phage with no side effects. Analysis of genetic similarities between phage wksl3 and four of its relatives (SS3e, vB_SenS-Ent1, SE2, and SETP3) allowed wksl3 to be categorized as a SETP3-like phage. A single-dose test of oral toxicity with BALB/c mice resulted in no abnormal clinical observations. Moreover, phage application to chicken skin at 8°C resulted in an about 2.5-log reduction in the number of Salmonella bacteria during the test period. The strong, stable lytic activity, the significant reduction of the number of S. Enteritidis bacteria after application to food, and the lack of clinical symptoms of this phage suggest that wksl3 may be a useful agent for the protection of foods against S. Enteritidis and S. Typhimurium contamination.

INTRODUCTION

Each year, nontyphoid Salmonella is involved in approximately 1.0 million estimated cases of salmonellosis, with 19,336 hospitalizations and 378 deaths, in the United States (1). A recent report by the U.S. Centers for Disease Control and Prevention (CDC) chronicled a continuous increase in Salmonella outbreaks from May to September 2010 throughout the United States; in these outbreaks, 1,608 illnesses were found to be associated with the consumption of shell eggs contaminated with Salmonella enterica serovar Enteritidis (2). In the Republic of Korea, S. Enteritidis and S. Typhimurium have been the main Salmonella serovars responsible for diarrhea and food-borne diseases due to salmonellosis in humans (3). In addition, S. Enteritidis is the second-most-reported serovar that causes Salmonella-related human disease in the United States, while S. Typhimurium is the most prevalent Salmonella serovar that causes salmonellosis (4). The industry has been investigating an effective process for the production of S. Enteritidis-free eggs and chickens worldwide (5, 6). In addition to the increasing incidence of salmonellosis, antimicrobial resistance in isolates associated with clinical outbreaks or food samples from food-borne outbreaks has been recognized as a new risk factor for Salmonella infection (7–10). The use of antibiotics in food-producing animals raised the prevalence of antimicrobial-resistant bacteria, and they have had adverse effects on the health of consumers via the food chain. The relationship between food-borne pathogens of human and animal origins has been well studied (11).

Widespread antibiotic resistance in isolates from various sources has encouraged many researchers to investigate and research phages as alternative biocontrol agents (12, 13). The use of phages as biological agents to control pathogens in foods has recently been suggested (14, 15). The use of a six-listeriaphage mixture to surface treat ready-to-eat meat and poultry products was approved by the U.S. Food and Drug Administration (FDA) in 2006, and in 2007, the U.S. FDA gave a generally recognized as safe (GRAS) designation to Listeria phage P100 (GRAS notice GRN 000218) for all products; P100 had already been approved for use in ready-to-eat foods as a food additive (16). Recently, P100 was listed by the Organic Materials Review Institute as an organic material classified as a processing nonagricultural ingredient and processing aid (http://www.omri.org/manufacturers/66440/ebi-food-safety-bv). The European Food Safety Authority also confirmed the safety of phage P100 as an antibacterial agent against Listeria monocytogenes on the surface of raw fish (17).

The phage application field is now expanding to target various food-borne pathogens and food products. In addition to the phage application test against L. monocytogenes (18, 19), studies investigating various food-borne pathogens, such as Salmonella spp. (20, 21) and Escherichia coli O157:H7 (22), have shown that phages are useful tools for the control of pathogens in foods without the risk of side effects. Since the regulatory clearance of the E. coli O157:H7-specific phage in the form of a food contact notification (FCN), the product can now be applied to red meat (FCN no. 1018). Moreover, another product based on a Salmonella phage is currently under review for FCN approval (Intralytix, Baltimore, MD).

In this report, we describe the detailed characterization and genetic information of Salmonella-specific virulent phage wksl3. Additional analyses, including phage host spectrum testing against S. Enteritidis, S. Typhimurium, and other serotypes; antimicrobial resistance profiles; oral toxicity tests; bioinformatic analysis; and efficacy in the control of S. Enteritidis on chicken skin, were also carried out to evaluate the potential of wksl3 for use as a food additive.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

A total of 111 Salmonella strains were used in this study (Table 1). Isolates from various food and clinical samples, such as ready-to-eat foods, livestock, fruits, vegetables, and clinical fecal samples, were collected from 2002 to 2010. The first S. Enteritidis isolate (SAL111-CF-KF10) producing the CTX-M-15 extended-spectrum beta-lactamase, acquired from a pediatric patient suffering from gastroenteritis in 2008 (23), was kindly provided by W. K. Song, Hallym University College of Medicine. Antimicrobial resistance phenotypes are also indicated in Table 1. Salmonella strains were grown at 37°C in tryptic soy broth (Bacto TSB; BD, Sparks, MD) or Bacto TSB supplemented with 1.5% agar. All strains were stored at −80°C in skim milk.

Table 1.

Antimicrobial resistance profiles and phage susceptibilities of the Salmonella strains used in this study

Host strain	Sample origin	Serotype	Antibiotic resistancea		Phage wksl3 susceptibilityb
			Resistance	Intermediate
S. Enteritidis
SAL1-FF-KW02	Processed foods	Enteritidis	NA		++++
SAL2-FF-KW02	Processed foods	Enteritidis	NA	CF	+++
SAL5-FF-KW02	Processed foods	Enteritidis		S	+++
SAL6-FF-KW02	Processed foods	Enteritidis			++
SAL7-FF-KW02	Processed foods	Enteritidis	NA		++++
SAL10-FF-KW02	Processed foods	Enteritidis	NA		++++
SAL11-FF-KW02	Processed foods	Enteritidis	NA		++++
SAL13-FF-KW02	Processed foods	Enteritidis		NA	++
SAL16-FF-KW02	Processed foods	Enteritidis		NA	++
SAL17-FF-KW02	Processed foods	Enteritidis		S, NA	+++
SAL18-FF-KW02	Processed foods	Enteritidis			++
SAL19-FF-KW02	Processed foods	Enteritidis			++
SAL23-FF-KW02	Processed foods	Enteritidis			++
SAL24-FF-KW02	Processed foods	Enteritidis			+++
SAL28-FF-KW02	Processed foods	Enteritidis	AM	CF, S	+++
SAL29-FF-KW02	Processed foods	Enteritidis	AM		+++
SAL30-FC-KW02	Livestock	Enteritidis		NA	+++
SAL32-FC-KW02	Livestock	Enteritidis	NA		+++
SAL33-FC-KW02	Livestock	Enteritidis	NA		+++
SAL57-FC-KF04	Livestock	Enteritidis	NA		++++
SAL58-FC-KF04	Livestock	Enteritidis	NA		+++
SAL61-FC-KF04	Livestock	Enteritidis	AM, S, C, TE		++++
SAL62-FF-KF04	Processed foods	Enteritidis	AM, S, C		++++
SAL63-FF-KF04	Processed foods	Enteritidis	AM, S, C	S	++++
SAL65-FC-KF05	Livestock	Enteritidis	S, NA, TE		++++
SAL66-FC-KF05	Livestock	Enteritidis	NA		+
SAL67-FC-KF05	Livestock	Enteritidis	AM, S, C		++++
SAL74-UI-KK	Unidentified	Enteritidis		S	++++
SAL97-FF-KK09	Livestock	Enteritidis	AM, CF, S, NA, C	SXT	++++
SAL99-FC-KK09	Livestock	Enteritidis	NA	S	++++
SAL105-FC-KK09	Livestock	Enteritidis	AM, CF, CTX, S, GM, NA, TE		+++
SAL111-CF-HU10	Clinical feces	Enteritidis	AM, CF, CTX, NA, TE	FOX, SXT	+++
S. Typhimurium
SAL3-FF-KW02	Processed foods	Typhimurium		NA	+++
SAL4-FF-KW02	Processed foods	Typhimurium		S, NA	+++
SAL8-FF-KW02	Processed foods	Typhimurium		NA	+++
SAL9-FF-KW02	Processed foods	Typhimurium		C	+++
SAL12-FF-KW02	Processed foods	Typhimurium		NA	+++
SAL14-FF-KW02	Processed foods	Typhimurium			+++
SAL15-FF-KW02	Processed foods	Typhimurium			+++
SAL20-FF-KW02	Processed foods	Typhimurium		S, NA	+++
SAL21-FF-KW02	Processed foods	Typhimurium			+++
SAL22-FF-KW02	Processed foods	Typhimurium		S	+++
SAL25-FF-KW02	Processed foods	Typhimurium		S	+++
SAL26-FF-KW02	Processed foods	Typhimurium		S	+++
SAL27-FF-KW02	Processed foods	Typhimurium		S, NA	+++
SAL34-FV-KW02	Fruit-vegetables	Typhimurium		S	+++
SAL35-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL36-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL37-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL38-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL39-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL40-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL41-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL42-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL43-FE-KW02	Other foods	Typhimurium			+++
SAL44-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL45-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL46-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL47-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL48-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL49-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL50-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL51-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL52-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL53-FV-KW02	Fruit-vegetables	Typhimurium			+++
SAL54-FC-KW02	Livestock	Typhimurium		S, NA	+++
SAL56-FF-KW02	Processed foods	Typhimurium			+++
SAL72-UI-KK	Unidentified	Typhimurium	AM, S, C, TE	AmC	++
Other S. enterica serotypes
SAL101-FC-KK09	Livestock	Agona		S	+++
SAL70-EW-KK	Environments	Agona		S	+++
SAL71-FF-KK	Processed foods	Anatum		S	−
SAL78-EW-KU09	Environments	Arizonae	S	TE	−
SAL79-VS-KU09	Vet feces	Arizonae	AM, S, C, TE		−
SAL83-VS-KU10	Carcasses	Arizonae	AM, CF, S, TE	AmC	−
SAL84-VS-KU10	Carcasses	Arizonae	AM, CF, S, SXT, TE	AmC, GM, NA, CIP	−
SAL85-VS-KU10	Carcasses	Arizonae	AM, CF, C, TE	SXT	−
SAL86-VS-KU10	Carcasses	Arizonae	AM, CF, S, SXT, TE	AmC, GM, NA, CIP	−
SAL87-FP-KU10	Livestock	Arizonae	AM, CF, S, GM, C, SXT, TE		−
SAL94-EW-KU10	Environments	Arizonae	S, TE		−
SAL108-VS-KU10	Carcasses	Arizonae	AM, CF, S, NA, TE	AmC, CTX	−
SAL109-VS-KU10	Carcasses	Arizonae	AM, CF, S, NA, TE	AmC	−
SAL110-VS-KU10	Carcasses	Arizonae	AM, S, C, TE	AmC, CF, SXT	−
SAL103-FC-KK09	Livestock	Dessau	NA	S	−
SAL106-FC-KK09	Livestock	Dessau	GM, NA	S	−
SAL31-FC-KW02	Livestock	Haardt	AM, S, NA, TE	CF	−
SAL55-FC-KW02	Livestock	Haardt	AM, S, NA, TE	CF	−
SAL59-FC-KF04	Livestock	Haardt	AM, S, NA, TE		−
SAL60-FC-KF04	Livestock	Haardt	AM, S, GM, NA, TE		−
SAL73-UI-KK	Unidentified	Heidelberg	S, GM		++
SAL68-FF-KK	Processed foods	Infantis		S	−
SAL69-FE-KK	Other foods	Javiana		S	++++
SAL77-FM-KK	Marine foods	Kentucky		S	−
SAL95-FF-KK09	Livestock	London		S	−
SAL98-FF-KK09	Livestock	London		S	−
SAL76-FF-KL	Processed foods	Montevideo		S	−
SAL100-FC-KK09	Livestock	Montevideo	NA		−
SAL102-FC-KK09	Livestock	Montevideo	NA	S	−
SAL75-FM-KK	Marine foods	Poona	AmC	S	−
SAL64-FB-KF05	Livestock	Rissen	TE	S	−
SAL96-FF-KK09	Livestock	Rissen	S, C, TE		−
SAL104-FC-KK09	Livestock	Weltevreden			−
SAL80-CF-KU09	Clinical feces	Nontyped	AM, S, TE	GM	++
SAL81-CF-KU09	Clinical feces	Nontyped	AM, S, NA, C, TE		+
SAL82-CF-KU09	Clinical feces	Nontyped	AM, S, TE		+++
SAL88-EM-KU10	Environments	Nontyped		S	−
SAL89-EM-KU10	Environments	Nontyped		S	−
SAL90-EM-KU10	Environments	Nontyped		S	−
SAL91-EM-KU10	Environments	Nontyped		S	±
SAL92-EM-KU10	Environments	Nontyped		S	±
SAL93-EM-KU10	Environments	Nontyped		S	−
SAL107-FC-KK09	Livestock	Nontyped	GM, NA	S	−

^aAmpicillin, AM; amoxicillin-clavulanic acid, AmC; cephalothin, CF; cefoxitin, FOX; cefotaxime, CTX; streptomycin, S; gentamicin, GM; nalidixic acid, NA; ciprofloxacin, CIP; chloramphenicol, C; trimethoprim-sulfamethoxazole, SXT; tetracycline, TE.

^b++++, complete lysis with secondary infection; +++, complete lysis; ++, lysis; +, turbid lysis; ±, growth inhibition zone; −, no plaques.

Isolation of Salmonella phage.

To isolate a Salmonella phage, we collected 25 chicken by-product samples from 16 traditional markets in GyeongGi-do, Republic of Korea. Three-gram samples were soaked in 30 ml sodium chloride-magnesium sulfate (SM) buffer with gelatin (100 mM NaCl, 10 mM MgSO₄ [heptahydrate], 50 mM Tris-HCl [pH 7.5], 0.01% gelatin). The tubes were vigorously vortexed for at least 5 min at room temperature. After centrifugation of the suspension at 4,500 × g for 30 min, the supernatant was filtered through a 0.20-μm membrane filter (Advantec Co., Ltd., Saijo City, Ehime, Japan). One hundred microliters of filtrate from each sample was then added to 4 ml Luria-Bertani (LB) broth supplemented with 10 mM CaCl₂ and 40 μl of an overnight broth culture of S. Enteritidis ATCC 13076 as a propagating host. After overnight incubation of the phage-Salmonella mixture at 37°C, each culture was filtered (0.20-μm filter) and standard plaque assays were performed with an indicator host (ATCC 13076) for each filtrate. Phage purification was carried out by picking single plaques with sterilized pipette tips, followed by serial purifications with amplifications from the same host (ATCC 13076), as described previously (24).

Determination of the phage host spectrum.

To confirm the host lysis range of Salmonella phage wksl3, spotting assays performed by the modified bilayer standard plaque assay method (25) were used with 111 Salmonella strains. After overnight cultivation of the Salmonella host, 40 μl of each solution was added to 4 ml of 50°C LB soft agar (0.75%) containing CaCl₂ (final concentration, 10 mM) and then the mixture was poured onto 1.5% LB agar plates. After a 30-min drying procedure in a laminar-flow closet, 5 μl of the diluted phage solution (1.1 × 10⁸ PFU/ml) was spotted onto the plate and incubated overnight at 37°C. All tests were conducted at least three times with all of the strains used in this study.

Transmission electron microscopy (TEM).

A purified wksl3 solution (1.2 × 10¹¹ PFU/ml) was dropped onto a Formvar carbon-coated copper grid (200 mesh). After 30 s of immobilization, water was removed with filter paper and then 2% (wt/vol) uranyl acetate was dropped onto the grid for negative staining of the phage. Electron micrographs were taken with a Carl Zeiss LEO 912AB transmission electron microscope operating at an 80-kV accelerating voltage. Images were taken on the transmission electron microscope at the National Academy of Agricultural Science, Suwon, Republic of Korea.

One-step growth curve.

The standard one-step growth curve of phage wksl3 was determined at 37°C as described by Ellis and Delbruck (26). S. Enteritidis ATCC 13076 was used as the host, and wksl3 was inoculated at a multiplicity of infection (MOI) of 0.01. Growth curves were plotted with Sigma Plot 10.0 (SPSS Inc.) based on PFU data collected every 5 min. The latency period and burst size were then calculated.

Nucleotide extraction, sequencing, and genomic analysis.

A high-titer (1.1 × 10¹¹ PFU/ml) phage solution was prepared for wksl3 DNA extraction. Three milliliters of wksl3 suspension was incubated with 30 μl DNase I (10 μg/ml; Sigma-Aldrich, UK) and RNase A (10 μg/ml; Sigma-Aldrich, UK) at 37°C for 50 min. Following the addition of 0.5% sodium dodecyl sulfate (SDS) and 50 μg/ml proteinase K, the phage suspension was incubated at 56°C for 1 h. After DNA extraction by the alkaline lysis method, 3 M sodium acetate (pH 5.4, 0.1 volume) and cold ethanol (2.5 volumes) were added for DNA precipitation as previously described (27). The phage genome was sequenced by the shotgun full-sequencing strategy on a 454 Genome Sequencer FLX titanium sequencer (Roche, Mannheim, Germany) at Macrogen Inc., Seoul, Republic of Korea. The whole genome sequence was also assembled by Macrogen Inc. with SeqMan II sequence analysis software (DNASTAR).

Possible open reading frames (ORFs) were predicted with the genome annotation software GeneMarkS (28) and confirmed with FgenesV (SoftBerry) and Glimmer 3.02 (29) by submitting the whole genome of wksl3. ATG, GTG, and TTG were considered gene start codons. Putative functions of conserved protein domains were identified with alignment search tools (BLASTP, BLASTX, and BLASTN search) found at the National Center for Biotechnology Information (NCBI) database. The PFAM 26.0 online program was used to search specific protein domains with known functions (30). Potential promoter regions for the wksl3 genome sequence with cutoff scores of 0.90 and 0.95 were examined with the Neural Network Promoter Prediction program (31) of the Berkeley Drosophila Genome Project (http://www.fruitfly.org/seq_tools/promoter.html). FindTerm programs (Softberry, Inc., Mount Kisco, NY) were used to identify nonoverlapping rho-independent terminators (at an energy threshold value of −11). Computed molecular weights and isoelectric points (pIs) of wksl3 putative protein products were predicted with proteomic tools (32) from ExPASy (http://www.expasy.org/proteomics). The tRNAscan-SE 1.21 program (http://lowelab.ucsc.edu/tRNAscan-SE/) was used to search for putative tRNAs (33). Sequence homologies of SETP3-like phages (wksl3, SETP3, SS3e, SE2, and vB_SenS-Ent1) were measured with the BLAST2 sequence tool (http://www.ncbi.nlm.nih.gov/blast/bl2seq/wblast2.cgi). Comparisons of equivalent ORFs of wksl3, SETP3, and SS3e were also performed with the same software. TMHMM 2.0 was used to predict transmembrane helices in proteins (34). To examine the similarity between wksl3 protein products and putative protein food allergens, the Allergenic Protein Sequence Searches program (35) of the Food Allergy Research and Resource Program database was used.

Oral toxicity studies with mice.

Oral toxicity studies were conducted with eight male BALB/c mice (Koatech, GyeongGi-Do, Republic of Korea) 8 weeks old and weighing between 21 and 23 g each. An acute oral toxicity test was performed according to the Good Laboratory Practice Standards manual and Organization for Economic Cooperation and Development (OECD) Guidelines for Acute Toxicity of Chemicals no. 420 (36). Mice were housed in a temperature-controlled animal room on a 12-h light-dark cycle. Fresh water and food were provided ad libitum throughout the experimental period.

After an aliquot of wksl3 (1.1 × 10¹⁰ PFU/ml) suspended in SM buffer at pH 7.5 was prepared, test groups containing five animals were orally administered stock solutions according to body weight (1 ml/100 g body weight) as suggested by the OECD guidelines. The negative-control group received SM buffer only in the same ratio as the test group. Feeding was permitted 4 h after dosing.

Animals were weighed before the test started and 1 week after the test period. The development of abnormal behavior, changes in physical appearance, and any other toxicological effects was observed within the first 6 h after the test solution was administered.

Complete gross pathological examinations of the skin, lymph nodes, bladder, testes, stomach, intestines, cecum, colon, spleen, pancreas, liver, kidneys, heart, thymus, and oral cavity of all test animals were conducted. All animals were euthanized by carbon dioxide asphyxiation. The necropsy results of all test animals were compared with those of negative-control animals.

Application of phages to control Salmonella on chicken skin.

Chicken skin samples were collected from broiler carcasses to verify the effectiveness of wksl3 for the control of experimentally contaminated S. Enteritidis on the surface of chicken skin. Sample preparation was performed as previously described, with modifications (37), and a total of 50 skin pieces (4 cm²) were prepared.

SAL111-CF-KF10 (5 × 10⁴ CFU/ml; selective marker is nalidixic acid and cefotaxime resistance) (23) was sprayed onto chicken skin with a disposable hand sprayer that transferred 0.2 ml per operation to give an initial concentration of approximately 10³ to 10⁴ CFU/cm². Each skin piece was inoculated individually. Sprayed samples were dried under blowing air for 1 h at 8°C, which is the average temperature of a domestic refrigerator (38). Phage solutions (2.2 × 10⁸ PFU/ml) were diluted with phosphate-buffered saline (PBS) and inoculated at an MOI of approximately 5 × 10³ by the same spraying method as previously described. Phage solutions were applied to the chicken skin one at a time shortly after Salmonella contamination. Twenty-five skin pieces were treated with phage solutions, and the remaining skin pieces were treated with the same volume of PBS as a control. Each skin piece was homogenized and diluted in 4 ml PBS. Viable cells were counted immediately after phage administration (day 0) and on test days 1, 2, 3, 5, and 7 by using four pieces each day. To determine the number of viable strains per skin piece, LB agar plates containing 128 μg/ml nalidixic acid and 8 μg/ml cefotaxime were used as selection media.

Statistical analysis.

Differences in weight changes and the numbers of viable bacteria between phage-treated and untreated groups were statistically analyzed with paired t tests and Duncan's multiple-range tests. P values of less than 0.05 were considered statistically significant. All tests were conducted with IBM SPSS Statistics 20 software.

Nucleotide sequence accession number.

The complete genome sequence of virulent Salmonella phage wksl3 was deposited in GenBank under accession number JX202565.

RESULTS

Bacteriophage wksl3 isolation and host spectrum determination.

Ten bacteriophage types were recovered from chicken by-product samples with S. Enteritidis ATCC 13076 as an indicator strain. Spotting assays were performed with various Salmonella strains, and one broad host spectrum phage, designated wksl3, was selected for further analysis. Phage wksl3 showed 100% lysis activity against all of the S. Enteritidis (n = 32) and S. Typhimurium (n = 36) strains tested and also infected S. enterica serovars Agona, Heidelberg, and Javiana, as well as 68% of the Salmonella strains tested (111 isolates). Positive lytic reactions were observed regardless of the host's antimicrobial resistance, showing that both drug-sensitive and -resistant S. Enteritidis and S. Typhimurium were phage susceptible (Table 1).

Phage wksl3 microscopy.

Morphological analysis revealed that phage wksl3 belongs to the order Caudovirales and family Siphoviridae morphotype B1 (the isometric head was 63 nm, and the long. noncontractile tail was 121 by 7.9 nm, with a 20-nm-wide baseplate with tailspikes), similar to the morphology of phage SETP3 (62.5 nm, icosahedral). Tails were rigid and noncontractile, measured 120 by 7 nm, and exhibited a 20-nm-wide baseplate with spikes (39).

One-step growth curve.

More than 90% of the wksl3 particles attached to the host cell within the first 10 min. Phage wksl3 displayed a 19-min latency period with a calculated average burst size of 51 PFU/cell (data not shown).

DNA sequence analysis.

According to sequence analysis, the wksl3 genome is composed of 42,766 bp, including 133-bp direct repeats at the end of the genome, with a total G+C content of 49.81%, which is similar to that of another Siphoviridae phage, SETP3 (42,572 bp in length with a G+C content of 49.85%). A total of 54 putative promoters, 20 transcriptional termination regions, and 64 ORFs, representing 91.6% of the phage sequence, were predicted in its genome. Genes were located in a region of high density (1.501 genes/kb), and the average length of each gene was 620 bp. Many short overlapping regions between contiguous genes were commonly detected. No predicted tRNA genes were discovered by tRNAscan-SE software. On the basis of transcriptional direction, the genome was predicted to be clustered into four groups. Twenty-three gene products showed significant homology to reported functional genes. Two ORFs were found to be members of the helix-turn-helix (HTH) superfamily, and one ORF was found to be a member of the immunity to superinfection membrane superfamily. While 38 ORFs were shown to encode hypothetical proteins, one ORF in wksl3 (gp33) showed no obvious homology to any other bacterium-, phage-, or prophage-related genes in the current GenBank database.

Bioinformatic studies of all 64 gene products of phage wksl3 showed no similarities to any other known virulent, toxin, or pathogen-associated protein family or gene product of Salmonella or any other pathogenic bacterium. With a 0.01 E-value cutoff, none of the protein products from the 64 predicted wksl3 genes were matched with polypeptides or protein sequences contained in the food allergenic protein sequence database.

According to the homology search-based annotation of functional genes, the wksl3 genes were categorized into three functional groups (Fig. 1): cell wall lysis genes (gp1, putative amidase; gp14, lysozyme; gp15, putative holin), phage structural genes (gp2, structural protein; gp3, terminase; gp4, terminase small subunit; gp38, tailspike protein; gp39, tail protein; gp43, tape measure protein; gp51, 53, 54, 58, and 63, tail proteins; gp59, head protein; gp60, coat protein; gp64, head morphogenesis protein), and metabolism-related genes (gp25, helicase-primase; gp32, DNA polymerase; gp34, restriction endonuclease; gp36, helicase; gp48, HNH endonuclease, gp50, DNA-binding protein). ORF information, such as the positions of genes, amino acid lengths, directions of transcription, sizes, functions, and homologies between wksl3 genes and other phage-related genes, is shown in Table 2.

Fig 1.

Genomic structure of wksl3, including the terminal redundant region. Arrows represent predicted genes and indicate their directions of transcription. Different colors denote different functional groups of bacteriophage genes.

Table 2.

Full genome sequence of wksl3

ORF	wksl3 coordinates (amino acid length)	Strand	Molecular mass (kDa)	pI	Best match (%)	Function	Best E value
gp1	42539–559 (216)	+	24.66	6.45	SETP3 orf32 (73)	Putative amidase	3E-107
gp2	590–2065 (491)	−	53.92	4.71	SETP3 orf31 (94)	Putative structural protein	0
gp3	2078–3349 (423)	−	47.52	5.86	SETP3 orf30 (96)	Terminase	0
gp4	3339–3845 (168)	−	18.78	5.99	vB_SenS-Ent1 orf1 (98)	Putative terminase small subunit	1E-116
gp5	3874–4005 (43)	−	4.71	4.22	SE2 orf29 (86)		6E-18
gp6	4141–4248 (35)	−	3.93	4.56	SS3e orf18 (100)		3E-16
gp7	4245–4547 (100)	−	11.60	9.52	SS3e orf19 (100)		3E-64
gp8	4547–4771 (74)	−	8.58	9.34	SS3e orf20 (100)		6E-47
gp9	4768–4917 (49)	−	5.93	5.4	SS3e orf21 (98)		1E-23
gp10	4914–5144 (76)	−	8.87	5.62	SS3e orf22 (99)		4E-46
gp11	5141–5311 (56)	−	6.20	10.29	SS3e orf23 (100)		2E-32
gp12	5293–5451 (52)	−	6.00	4.3	vB_SenS-Ent1 orf54 (79)		4E-08
gp13	5448–5633 (61)	−	6.89	7.89	SETP3 orf24 (93)		3E-32
gp14	5820–6308 (162)	−	17.50	9.71	SETP3 orf23 (93)	Lysozyme	3E-97
gp15	6286–6576 (96)	−	10.61	9.8	vB_SenS-Ent1 orf51 (83)	Putative holin	9E-52
gp16	6578–6859 (93)	−	10.29	7.87	vB_SenS-Ent1 orf50 (98)		6E-60
gp17	6939–7373 (144)	−	15.38	4.63	SS3e orf27 (100)		7E-101
gp18	7379–7744 (121)	−	13.82	10.52	SS3e orf28 (100)		1E-82
gp19	7741–7944 (67)	−	7.32	7.79	SS3e orf29 (100)		6E-41
gp20	7941–8546 (201)	−	23.54	5.73	SS3e orf30 (100)		1E-123
gp21	8539–8652 (37)	−	4.10	3.4	SS3e orf31 (100)		3E-16
gp22	8715–8879 (54)	−	6.41	11.57	SETP3 orf18 (89)		1E-26
gp23	9736–9921 (61)	+	7.28	8.84	SS3e orf32 (100)	DNA-binding helix-turn-helix superfamily protein	9E-37
gp24	9918–10151 (77)	+	8.60	9.61	SS3e orf33 (99)		2E-46
gp25	10208–12394 (728)	+	80.57	4.97	vB_SenS-Ent1 orf43 (98)	Putative replicative helicase-primase	0
gp26	12386–12628 (80)	−	8.84	8.8	SETP3 orf14 (100)	Helix-turn-helix family protein	2E-44
gp27	12762–13274 (170)	+	19.16	4.65	SE2 orf11 (99)		2E-112
gp28	13318–13521 (67)	+	7.66	7.87	SE2 orf10 (100)		1E-41
gp29	13518–13790 (90)	+	10.78	9.2	SE2 orf9 (100)		8E-62
gp30	13787–15061 (424)	+	47.16	6.35	SE2 orf8 (99)		0
gp31	15143–15772 (209)	+	23.63	4.85	SE2 orf7 (98)		1E-134
gp32	15830–18031 (733)	+	82.51	6.93	SETP3 orf10 (98)	DNA polymerase	0
gp33	18119–18271 (50)	+	5.42	7.81
gp34	18268–18552 (94)	+	10.82	9.56	vB_SenS-Ent1 orf36 (96)	Putative restriction-endonuclease	9E-58
gp35	18584–18775 (63)	+	7.03	9.45	vB_SenS-Ent1 orf35 (100)		1E-36
gp36	18772–21237 (821)	+	92.29	8.87	SETP3 orf6 (98)	Putative helicase	0
gp37	21234–21455 (73)	+	8.72	4.85	SS3e orf48 (100)		1E-41
gp38	21574–23604 (676)	−	72.11	5.22	SETP3 orf4 (95)	Tailspike protein	0
gp39	23641–26127 (828)	−	90.96	5.18	vB_SenS-Ent1 orf31 (98)	Putative tail protein	0
gp40	26190–26453 (87)	−	9.94	5.41	vB_SenS-Ent1 orf30 (95)		1E-54
gp41	26552–27067 (171)	−	19.15	4.46	SE2 orf58 (99)		3E-120
gp42	27064–27564 (166)	−	16.49	4.59	SS3e orf54 (98)		2E-118
gp43	27566–29899 (777)	−	83.14	4.75	vB_SenS-Ent1 orf27 (98)	Putative tape measure protein	0
gp44	29892–30251 (119)	−	13.61	4.69	SE2 orf55 (97)		2E-81
gp45	30257–30673 (138)	−	15.87	5.28	vB_SenS-Ent1 orf25 (99)		6E-96
gp46	30843–31022 (59)	+	6.60	9.7	SETP3 orf50 (100)	Immunity to superinfection membrane superfamily protein	3E-32
gp47	31085–32203 (372)	+	42.22	8.66	SETP3 orf49 (99)		0
gp48	32256–32756 (166)	+	19.28	9.92	vB_SenS-Ent1 orf22 (99)	Putative HNH endonuclease	5E-117
gp49	32766–32951 (61)	+	6.89	6.01	SETP3 orf48 (82)		2E-21
gp50	33112–33783 (223)	+	25.63	6.24	vB_SenS-Ent1 orf20 (88)	Putative DNA-binding protein	3E-138
gp51	33812–34981 (389)	−	41.20	4.64	vB_SenS-Ent1 orf19 (99)	Putative tail protein	0
gp52	34981–35400 (139)	−	15.06	4.56	SETP3 orf45 (100)		9E-97
gp53	35400–35795 (131)	−	14.47	9.75	SETP3 orf44 (97)	Putative tail protein	2E-88
gp54	35792–36151 (119)	−	12.98	9.4	vB_SenS-Ent1 orf16 (96)	Putative tail protein	71E-76
gp55	36151–36756 (201)	−	20.69	6.59	SE2 orf44 (94)		9E-131
gp56	36759–37268 (169)	−	17.81	4.84	SE2 orf43 (95)		4E-113
gp57	37272–37460 (62)	−	7.15	4.92	vB_SenS-Ent1 orf13 (98)		3E-35
gp58	37497–37847 (116)	−	12.23	4.38	vB_SenS-Ent1 orf12 (95)	Putative Hoc protein	3E-72
gp59	37859–38143 (94)	−	9.37	9.77	SETP3 orf38 (96)	Putative head protein	1E-46
gp60	38204–39253 (349)	−	37.89	4.84	SETP3 orf37 (97)	Putative coat protein	0
gp61	39257–39958 (233)	−	25.72	5.63	SS3e orf10 (99)		2E-162
gp62	40149–40538 (129)	−	14.39	9.27	SS3e orf11 (100)		4E-89
gp63	40856–41314 (152)	−	16.40	4.59	SETP3 orf34 (95)	Tail protein	1E-95
gp64	41317–42360 (347)	−	38.58	6.39	SETP3 orf33 (96)	Head morphogenesis protein	0

Dot plot analysis of wksl3 and other Salmonella phages (SETP3, SS3e, vB_SenS-Ent1, and SE2) revealed that they show significant homology at the nucleotide sequence level (Fig. 2B), suggesting that these phages can be clustered into a tentative relational group. BLAST2 sequence comparisons with other SETP3-like phages revealed that the full wksl3 genome sequence shows 84% homology to that of SETP3 (GenBank accession number EF177456), 89% with that of vB_SenS-Ent1 (GenBank accession number HE775250), 90% with that of SE2 (GenBank accession number JQ007353), and 93% with that of SS3e (GenBank accession number AY730274). CoreGenes 3.0 analysis, with the BLASTP threshold score set at 75, revealed that these five phages have a total of 40 homologous genes in common, while wksl3 has 50, 53, 56, and 54 homologous genes in common with SETP3, SS3e, SE2, and vB_SenS-Ent1, respectively (40). To compare the homologies of the ORFs in wksl3, SS3e, and SETP3, we used BLASTN and BLASTP analyses and confirmed the colinear and reverse complemented arrangement of functionally homologous genes or orthologous proteins (Fig. 2A), suggesting that wksl3 has the same ancestry as other SETP3-like phages, including SETP3, vB_SenS-Ent1, SE2, and SS3e.

Fig 2.

(A) Three homologous SETP3-like Salmonella phages. Rectangles represent the genes of each phage, and the numbers above the rectangles are ORF numbers. Numbers in parentheses under the rectangles indicate the corresponding wksl3 gene numbers that show high sequence homology. The color of each rectangle represents the functional group of each gene, predicted with BLAST programs. The scale is in kilobase pairs. (B) Dot plot image displaying sequence homology comparisons of wksl3 and four SETP3-like phages (SETP3, SS3e, vB_SenS-Ent1, and SE2).

Like other lytic SETP3-like phages, no proteins encoded by wksl3 genes showed homology to other phage lysogenic-cycle-related genes, such as integrases, in GenBank. In addition, PHACTS (41) results predicted that phage wksl3 is a lytic phage with a PHACTS output value of 0.532, corresponding to the fraction of trees in the algorithm supporting this conclusion. In fact, SETP3 and SE2 have also been reported as lytic (42, 43). This value indicates that wksl3 also has strong lytic activity; in comparison, the well-known, strongly virulent Listeria phage P100 (GenBank accession number DQ004855) was predicted to be lytic with a 0.525 value.

As demonstrated in Fig. 1, wksl3 genes were arranged into four transcriptional clusters representing the regulator operon (gp46 to gp50), the replication operon (gp23 to gp37), and two late operons. Structural genes and cell wall lysis-related genes were found in the late operons. A relatively large gene-free region (857 bp) was observed between gp22 and gp23, and this region was also observed in vB_SenS-Ent1 (957 bp) and SE2 (994 bp), located between genes orthologous to wksl3 gp22 and gp23.

The two subunits of gp1 of wksl3 were located at the right and left ends of the linear genome structure. Both subunits appeared to be transcribed in the same direction. Since circularly permuted genomes are commonly observed in head-full DNA packaging schemes (44), it is possible that the two gp1 subunits, which were separated by the terminal redundant region, may be translated as one single gene. The gp1 orthologs SE2 gp33 and SETP3 gp32 also showed single-gene translation. Moreover, gp1 exhibited significantly better homology to SE2 gp33 and SETP3 gp32 in terms of amino acid length and sequence identity than the combined or separate gene products of the two subunits.

Metabolism-related operons.

The phage wksl3 regulator operon is composed of five genes (gp46 to gp50) and is located between two structural gene clusters. gp50 shows high homology to phage vB_SenS-Ent1 gp20 (88%) and SETP3 gp47 (87%), each of which encodes a DNA-binding protein. These three orthologous genes also contained both the phage regulatory protein Rha (pRha) domain and the phage antirepressor protein KilAC domain (ANT). Phage SE2 gp49 showed significantly high sequence homology (99%) with gp20 but differed in that it lacked an N-terminal pRha domain. The P22 antirepressor ant encodes an antirepressor protein that binds to and inhibits the CII repressor protein of phage P22 (45). In addition, P22 ant activity against lambda repressor CI has also been observed (46). We expect that the antirepressor domain of wksl3 will function in lytic-activity development, similar to the P22 antirepressor protein (47). An HNH endonuclease was predicted to be encoded by gp48, which showed 99% homology to vB_SenS-Ent1 gp22 and 62% homology to phage T1 endonuclease, whereas no homologous gene was identified in the SETP3 genome. gp46 was predicted to possess immunity to superinfections via three predicted transmembrane domains (TMHMM), and this gene showed not sequence but functional homology to the vB_SenS-Ent1 immunity protein (gp24).

The replication operon (gp23 to gp37) of wksl3 encodes a replicative helicase-primase, a helicase, a DNA polymerase, a restriction endonuclease, two putative HTH DNA-binding regions, and nine hypothetical proteins. gp25 encodes a replicative helicase-primase containing an AAA_25 (pfam13481) domain and a hexameric replicative helicase RepA region. gp36 encodes a helicase consisting of a homing endonuclease intein domain of 346 amino acids (pfam05203) and an SNF2 family N-terminal domain. A BLASTP search of the wksl3 helicase revealed 98% sequence homology to the SETP3 helicase and the vB_SenS-Ent1 intein-containing helicase precursor. Without an endonuclease domain, gp36 showed significant identity with SE2 gp1 (99%) and SS3e gp47 (97%). The DNA polymerase encoded by gp32 showed homology to the SETP3 DNA polymerase (98%) and the vB_SenS-Ent1 intein-containing DNA polymerase precursor (98%), while the wksl3 DNA polymerase lacked a 299-amino-acid intein domain. gp34, which encodes a restrictive endonuclease, showed high similarity to the putative restriction endonuclease (96%) of vB_SenS-Ent1, gp44 (92%) of SS3e, and gp8 (89%) of SETP3. gp26 contains a Cro/C1-type HTH DNA-binding domain with a transcriptional direction that is the opposite of that of the replication operon, suggesting that gp26 may act as a regulator protein. gp23 also contains an HTH transcription regulator MerR superfamily region.

Structure- and cell wall lysis-related operons.

Two late operons (gp38 to gp45 and gp51 to gp64/gp1 to gp22) contained genes involved in phage structure and host cell lysis. The arrangement of genes encoding wksl3 structure assembly proteins followed the conserved synteny and gene orders of Siphovirus (48). Entire gene products encoded by the wksl3 late operon had orthologous gene products in other SETP3-like phages. PSI-BLAST and Pfam analysis results against a putative phage structural protein (gp2) suggested that the gene product immediately downstream of the two terminase subunits (gp3 and gp4) was predicted to be a portal protein. The major tail protein of wksl3 (gp51) showed 99% homology to the vB_SenS-Ent1 major tail protein and exhibited 99% and 93% similarity to the SE2 and SETP3 tail proteins, respectively. Comparisons of the tape measure protein lengths and TEM-based tail lengths of wksl3 and SETP3 confirmed that the tail length corresponded to the length of the tape measure protein, similar to other Siphoviridae phages (49, 50). BLASTP analysis of gp39 showed significant homology to the tail fiber proteins of vB_SenS-Ent1 (98%) and SE2 gp60 (98%). Interestingly, wksl3 gp39 also showed sequence similarity to putative tail proteins of Escherichia phages K1ind3 (62%), K1ind1 (61%), K1dep1 (61%), and K1dep4 (62%). gp38 encodes a tailspike protein that contains a phage P22 tailspike domain in the C-terminal region. The C-terminal sequence of the P22 tailspike protein, which possesses oligosaccharide-binding and endorhamnosidase activities, is crucial for binding to the O-antigen region of the bacterial lipopolysaccharide (51). These results indicate that wksl3 may be able to recognize lipopolysaccharide (LPS) O-antigenic repeating units to infect host cells. O-antigen affinity was also observed in phage SETP3 (39). BLASTP results showed significantly high homology to the tailspike proteins of Siphoviridae phages SE2 (98%), SETP13 (97%), SS3e (96%), vB_SenS-Ent1 (96%), SETP3 (95%), SETP7 (94%), SETP5 (95%), and SETP12 (95%). High local similarity to the P22 tailspike region in wksl3 gp38 was also found in Salmonella Podoviridae phages ST104 (80%), P22 (80%), SETP14 (80%), SPN9CC (79%), SE1 (79%), ST64T (79%), SETP1 (79%), and 15 (79%).

Phage wksl3 was expected to encode a holin (gp15), a lysozyme (gp14), and an amidase (gp1), which would affect the phage's mechanism of host cell wall lysis. In wksl3, lysis-related genes were located in the late operon. gp1 showed 76% and 73% C-terminal homology to the amidase of phages SE2 and SETP3. As in many other double-stranded DNA phages, gp15, a predicted holin composed of 96 amino acids with three helical transmembrane regions (34), and gp14, which encodes a lysozyme with 93% homology to the SETP3 lysozyme, seemed to be involved in the holin-endolysin system. Regarding the lysis cassette of the dual-protein system (52), up to five proteins (gp9 to gp13) encoded by the cassette were not assigned to any other auxiliary lysis functions. This system differs from that of Salmonella phage SE2 (42), which was predicted to use a holin-independent lytic system. Considering the predicted promoter and terminator regions, as well as the direction of gene transcription, the amidase and dual-lysin systems seemed to act independently on host cell lysis.

Acute oral toxicity study with mice.

No mice in either the test or the control group died during the 8-day study period. No clinical signs of wksl3-mediated toxicity were observed in the test group to which a single dose of 10¹¹ PFU/kg body weight was administered. No development of abnormal behavior, changes in physical appearance (such as hair loss or wound formation), or any other toxicological effects, including inflammation, allergy, or diarrheal symptoms, were observed in any mouse during the experimental period (data not shown). The mean body weights of the phage-treated and control groups before treatment did not differ significantly. However, average increases of 3.62 g (phage-treated group) and 2.77 g (phage-untreated group) were observed by day 8 and these changes were significantly different from the baseline measurements (P < 0.05). Moreover, the absolute weights of the mice in the two groups differed significantly as a result of these changes in weight (P < 0.05), with phage-treated mice showing a greater increase in body weight.

Macroscopic examination of the organs of all of the phage-treated mice revealed normal color compared to those of untreated animals (data not shown). Moreover, no obvious differences or gross lesions were found in the skin, lymph nodes, bladder, testes, stomach, intestines, cecum, colon, spleen, pancreas, liver, kidneys, heart, thymus, or oral cavity in either treated or untreated mice at postmortem examination. Thin sections were not observed in the gastrointestinal tracts of phage-treated mice. Autopsy results of stomach tissues from wksl3-treated mice showed that no adverse inflammatory effects occurred at either the outer or the inner surface of the stomach (data not shown). All examinations were performed and confirmed by a doctor of veterinary medicine (DVM). According to the DVM's opinion of their gross pathology, there were no noticeable abnormalities in phage-treated mice, supporting the conclusion that phage wksl3 is nontoxic.

Effectiveness of wksl3 in the control of Salmonella on chicken skin.

Figure 3 demonstrates the effects of phage wksl3 on Salmonella strains used to inoculate chicken skin. Skin pieces were experimentally sprayed with approximately log₁₀ 3.25 CFU of Salmonella Enteritidis/cm² of skin. A single-dose application of phage wksl3 resulted in a 3.04-log decrease (P < 0.001) in the numbers of viable Salmonella bacteria after 24 h of storage at 8°C. Viable Salmonella cell concentrations were reduced by log₁₀ 0.3 CFU/cm² (mean, 2 CFU/cm²) at day 1, but growth resumed after day 2. According to our statistical analysis, no significant growth was observed after days 2 to 7 (P = 0.13). Phage wksl3-treated samples maintained a decrease in the average viable Salmonella level of log₁₀ 2.43 CFU/cm² (P < 0.01) from day 2 to day 7. Forty-eight randomly selected viable cells recovered from wksl3-treated chicken skins after day 7 retained their susceptibility to phage wksl3.

Fig 3.

Effect of phage wksl3 on growth of S. Enteritidis on chicken skin incubated at 8°C. Diluted Salmonella cells (10³ CFU/cm² skin) were applied at day 0. Phage wksl3 (1 × 10⁷ PFU/ml, open circles) and buffered peptone water (negative control, closed circles) were applied after 2 h of stabilization at 4°C.

DISCUSSION

Bacteriophages (or, more simply, phages) are viruses that infect and kill bacterial cells. Generally, phages are found near the host bacteria (53) and recognize specific bacterial hosts (54). While the emergence of pathogenic bacteria with increasing resistance to currently used antimicrobial agents is growing, phages have now resurfaced as potential biocontrol and therapeutic agents (55–60). The majority of human-infecting Salmonella strains originate from poultry and poultry-derived food products (61). On this basis, we screened and isolated Salmonella-specific phages from chicken carcasses and by-products purchased from local markets in the Republic of Korea. A wide-host-spectrum virulent phage, wksl3, was selected and characterized to investigate its host spectrum in relation to antibiotic resistance patterns, genomic characterization, oral toxicity, and efficacy of model phage application to processed chicken carcasses from chicken processing plants for assessment of its adequacy as a potential biocontrol agent.

Phage wksl3 was obtained with S. Enteritidis ATCC 13076 as an indicator strain; the phage had strong lytic activity against S. Typhimurium and S. Enteritidis and also inhibited the growth of S. Agona, S. Heidelberg, and S. Javiana. Phage wksl3 formed different plaque turbidities with variable efficiencies of plating, regardless of the antimicrobial resistance patterns of the strains (Table 1). While S. Typhimurium and S. Enteritidis have been proven to be the major serotypes responsible for the majority of Salmonella-related outbreaks, this result implied that application of this phage could be an effective tool to reduce S. Enteritidis- and S. Typhimurium-related food-borne salmonellosis by inactivating the growth of S. Enteritidis and S. Typhimurium, including a few additional serotypes from the field. We were able to classify Salmonella phage wksl3 in the order Caudovirales and the family Siphoviridae (Fig. 4) on the basis of its morphology (an icosahedral head with a noncontractile tail) (62).

Fig 4.

Transmission electron micrograph of lytic bacteriophage wksl3. Bar, 200 nm.

In silico full-genome sequence analysis revealed that none of the wksl3 gene products showed any similarity to known pathogenic-bacterium-related toxin, pathogenic, or virulence-encoding genes. Bioinformatics also failed to indicate homology between any of the wksl3 gene products and potential immunoreactive food allergens. High homology in terms of genome size and sequence identity suggested that wksl3 is related to SETP3, SE2, vB_SenS-Ent1, and SS3e (39, 42, 63). In particular, we identified wksl3 as a novel member of the group of SETP3-like phages that is classified in the family Siphoviridae in an International Committee on Taxonomy of Viruses report released in 2011 (63) and in the NCBI taxonomy.

Bacteriophages are known to recognize various outer cell wall components, including flagella (64–66), LPS (67–70), and several outer membrane proteins (71–74). While phage cocktails are generally composed of phages with a spectrum of two or more hosts using different receptors, receptor identification is the first step in the measurement and evaluation of their potential applications in the biocontrol of target bacteria (75) and is useful for producing more efficient phage cocktails. According to our current genome analysis, the tailspike protein of wksl3 exhibits a high level of homology to various Salmonella phages, including SETP3-like phages. Tailspike proteins encoded by SETP3-like phages contain a phage P22 tail-spike domain in the C-terminal region. Both phages P22 and SETP3 are known to recognize the O antigen as a receptor (39, 51). Phage wksl3 encodes two endonucleases that play important roles in phage replication. Nuclear disruption of host DNA occurs because of the action of endonucleases, and disrupted nucleotide particles are recombined into progeny phage DNA (76). Methyltransferases or methylases in phages protect DNA from self or host cell restriction enzymes (77). The DNA methylase N-4/N-6 domain protein was encoded in SE2 (42), while other SETP3-like phages did not have homologous genes or putative methylase functional genes in their genomes. However, wksl3 gp35 showed 30 and 50% homology to Anaeromyxobacter dehalogenans 2CP-C lysine N-methyltransferase and Planctomyces brasiliensis DSM5305 guanine-N(2)-methyltransferase, respectively. gp47 also contained a domain in the N-terminal region similar to the Saccharomonospora cyanea NA-134 N-6 DNA methylase domain, with 28% similarity. gp35 and gp37 of phage wksl3 can be predicted to function as potential methylases and/or methyltransferases.

In our oral acute-toxicology study with mice to observe phage wksl3-related toxicity effects, a high-titer (1.1 × 10¹¹ PFU/kg body weight) single dose of phage wksl3 produced no test substance-related clinical signs, gross lesions, or deaths. The titer used in this study corresponds to 6.82 × 10¹² phage particles per average human body weight (62 kg) (78). According to a previous study (79), the total surface area of a 1.5-kg chicken carcass corresponds to 1,940 cm² and an uncooked chicken carcass will contain approximately 3.1 × 10¹⁰ phage particles. More than 200 chickens would have to be consumed by an individual to equal the phage titer tested. Considering global chicken and poultry consumption per year (80), this value is quite high. Previous studies have also reported that no deaths, abnormalities, or adverse effects were observed after the administration of phages to animals (18) and that oral administration of phages to humans is also safe, with no adverse effects (81). Advanced toxicity testing, including long-term toxicity tests and histopathological analysis, is required to confirm the safety of phage wksl3.

Phage wksl3 was first applied to LB broth at different MOIs and temperatures (22 and 8°C; data no shown). Phage application (1.1 × 10¹¹ PFU/ml) with overnight culture of the host strain at low temperatures resulted in complete elimination of Salmonella cells from the broth. Tests performed at 22°C also showed a significantly large (4- to 5-log) reduction but not complete eradication of viable Salmonella cells on day 1. However, phage-treated samples incubated at 8°C showed complete elimination of host cells over the 1-week test period. Hence, the present study demonstrated the optimal application conditions for phage particles on artificially contaminated chicken skin (low temperature and appropriate MOI). The Salmonella count was successfully reduced below the detection limit (30 CFU/ml). Regarding the phage susceptibility of the remaining strains in Fig. 3, escape from contact with phage particles because of their immobilization on food seems to cause incomplete Salmonella reduction. A standardized phage amount per unit of meat surface area and a method of homogeneous distribution on food are necessary to achieve highly efficacious pathogen reduction (19). For effective application of wksl3 to foods, the time points, frequency, and dosage of the phage on different types (forms, textures, etc.) of foods should also be optimized.

In a previous investigation at a chicken slaughtering plant, phage application in chiller water showed more efficient reduction (J. W. Kim et al., unpublished data), so it would be useful to test wksl3 in a processing plant in a future study.

The emergence of multiple-antibiotic-resistant pathogens has sparked interest in alternative antibiotics. Phages are predators of specific pathogens, and many Western countries have begun to investigate whether this natural bacterial enemy may be able to control pathogens that cannot be controlled by available drugs (82). Phages have merits in various contexts; (i) phage-resistant strains are still susceptible to other phages targeting different receptors and (ii) isolation of novel phages requires relatively less time than the development of new antibiotics (83). The most interesting finding in this study is that Salmonella phage wksl3 showed a bactericidal effect against all of the S. Enteritidis and S. Typhimurium strains tested, regardless of their antimicrobial resistance. This may suggest that this phage could be useful in the control of S. Enteritidis and S. Typhimurium, including antimicrobial-resistant strains, and in the protection of foods from bacterial contamination as a food preservative. In addition, these characteristics of phages suggest that further investigation of this natural predator may give us many advantages in the war against pathogens. Additional bioinformatic analyses with genome sequences of applicable phages should be conducted to provide us with information to evaluate the potential usefulness of phages or risk factors associated with phage consumption. Combined use of phages and other agents, such as bacteriocins or essential oils, may also be advantageous.

In conclusion, this study was carried out to characterize and analyze the novel SETP3-like phage wksl3. The strong lytic activity and broad host spectrum of phage wksl3 indicate that it has potential as a biocontrol agent against S. Enteritidis and S. Typhimurium. Our data indicate that this phage may be effective in S. Enteritidis reduction strategies, regardless of drug resistance, to ensure food safety and public health protection. In addition, no clear harmful effects were noted after oral administration to mice. This encouraging result implies that wksl3 may be a potential therapeutic agent not only for the prevention of contamination of foods but also for the treatment of humans or animals in clinical contexts. Further model application studies are needed to commercialize and expand the use of phage functions to other types of food products. Moreover, further genomic, proteomic, and host mutational studies of phage wksl3 may provide a greater understanding of the potential ecological role of wksl3 homologues in phage-host interactions during farm-to-table development, extending the range of phage applications.