Identification and Characterization of Salmonella Phages and Absence of Salmonella Strains from Three Different Study Areas of Cattle Livestock

Abstract

Background:

Salmonella represents a significant risk to both public and animal health. Although Salmonella strains have not been isolated from cattle livestock, Salmonella phages have been successfully identified.

Materials and Methods:

Our study was conducted through (i) investigating the presence of Salmonella strains and Salmonella phages in three study areas, (ii) phenotypic screening by lytic profile (LP); (iii) selecting nine phages for sequencing; and (iv) genomic comparison to evaluate their relative diversity.

Results:

A total of 307 samples were analyzed, resulting in a total of 162 virus-like particles (VLPs) analyzed. The LP was performed to identify Salmonella strains susceptible to phage infection, including the most frequent serovars: Dublin, Enteritidis, and Javiana. From the VLPs, nine phages were selected for genomic comparison. These phages represent three morphotypes: siphoviruses, myoviruses, and podoviruses, originating from different geographic and productive sites.

Conclusions:

This study enhances the understanding of the presence and diversity of Salmonella phages in cattle livestock, even in the absence of Salmonella strains.

Introduction

Salmonella is a foodborne pathogen affecting most animal species and humans and is considered a significant public health concern.¹ Cattle can acquire Salmonella enterica early and become ill with gastroenteritis symptoms, including intense diarrhea.² Shortly after infection, Salmonella can remain in the lymph nodes or be acquired during the animal’s life, which favors its shedding and dissemination. Different Salmonella serovars have been isolated from cattle lymph nodes,^3,4 carcasses, intestine, and adipose tissue^5,6; all of which might cause other symptoms such as pneumonia, septicemia, and reproductive losses.² In cattle, Salmonella can be transmitted directly through contact with bovine fluids, secretions, fecal material or indirectly through contaminated food, including meat carcasses and milk, or water sources.² Notably, characteristics of nontyphoidal Salmonella serovars also produce a substantial health burden in people, including poorer well-being and mortality, mainly attributed to invasive diseases.⁷ A global systematic review and meta-analysis have reported cattle-associated serovars resulting in human salmonellosis cases such as multidrug resistance of Salmonella Typhimurium lineage DT104,^8,9 and Salmonella Dublin is the most frequently isolated serovar from cattle worldwide, which can be considered endemic in heifer-rearing facilities.^10,11 Therefore, the circulation of Salmonella represents a substantial risk to animal and public health due to its abundant prevalence within animal systems.¹¹

Bacteriophages, or phages, are viruses capable of infecting and replicating in metabolically active bacteria and subsequently eliminating it (lytic effect).⁶ Due to the abundance of these agents in various ecological niches (both environments and animals) and the absence of a unique marker for their identification (e.g., 16S rRNA in bacteria), the taxonomic identification and description of their diversity is complex. The International Committee on Taxonomy of Viruses (ICTV)¹² currently recommends that taxonomic identification should include bioinformatic analysis of virus sequences, morphology of virus-like particles (VLPs), and information on their relatedness and observed biological properties (phenotypes).¹²

On the contrary, phages also offer an alternative for monitoring the presence of Salmonella and as a control agent for the pathogen in animals, especially those that are susceptible or immunosuppressed. Several authors have demonstrated the persistence of Salmonella phages in dairy environments, including industrial dairies in the United States,¹³ and small- to medium-sized dairies in Chile¹⁴ and Thailand.¹⁵ It is also interesting to note that in Chile there are antecedents of Salmonella phage circulation in animal systems^14,16; however, the isolation and characterization of Salmonella strains has not been proven in cattle-livestock.^14,16

To understand the characteristics that influence the maintenance of phage circulation in each dairy, it is important to point out that the classification of the size of farm (small or backyard, medium and large) is related to the economic resources for the expenses of biosecurity and animal maintenance.¹⁷ In this context, one simple and effective tool for classifying circulating phages in cattle-livestock is the lysis profile (LP) or host range.¹⁸ The LP allows the phenotypic classification of each VLP, according to the ability of infecting different strains of the hosts^14,18–21 because the lytic phages allow rapid detection of bacterial pathogens, due to the fast replication and short infection cycle time, together with the fact that they multiply in viable host cells with a low false positive rate with respect to other methods such as Polimerase Chain Reaction (PCR). So, phages can identify potential circulating Salmonella hosts.¹⁹ Previously, Salmonella phages have been grouped phylogenetically into a single cluster and distinct infection profiles.^19–21 Therefore, our study aimed to analyze the presence of phages and their phenotypic and genetic diversity in cattle with variable geographic and productive characteristics and a history of Salmonella absence.

Materials and Methods

Study area

This study was conducted in three different areas in Chile: (i) Easter Island; (ii) central Chile, and (iii) southern Chile (phage isolated previously). All production cattle-livestock in each selected area had a history of no isolation of Salmonella strains. So, we used nonprobabilistic convenience sampling,²⁰ and each sampling was identified using ArcGIS 10 software (Esri, Redlands, CA) with the coordinates obtained at each sampling time. Each sample area was georeferenced using a Geographical Positioning System, Gpsmap62s (Garmin, Oalthe, Kansas).

In detail, Easter Island is part of the Chilean insular territory and part of the Polynesian Triangle, it has subtropical weather, warm temperatures with rain all year round.²² This territory has a limited exploration in terms of pathogen transmission and is particularly noteworthy.²³ The productive destiny of the cattle-livestock is to obtain dairy products and meat, for self-supply, family consumption^17,20,23 classified them as backyard cattle-livestock (Fig. 1). We collected 47 samples from 90 cattle from six sites on Easter Island (Orito, Maunga Otu’u, Vaihu, Rano Raraku, Puna Pau, and Akahanga) (Fig. 1 and Table 1). Elsewhere, the central area, represented by Libertador General Bernardo O’Higgins (LGBO) region, has a Mediterranean climate (average temperature of 15°C and annual rainfall of 769 mm).²² This region is devoted to agriculture, with a few farms dedicated to small and medium cattle-livestock from these farms were taken 260 samples, collected from six farms (Table 1).

FIG. 1.

This study was carried out in three study areas, shown in the map, in green (A) Easter Island, Orito (farm 1), Maunga Otu’u (farm 2), Vaihu (farm 3), Rano Raraku (farm 4), Puna Pau (farm 5), and Akahanga (farm 6); (B) in blue central Chile, General Libertador Bernardo O’Higgins region, Rengo (farms 2 and 3), Malloa (farms 5, 4, and 11), Quinta de Tilcoco (farms 6 and 1), Chimbarongo (farms 7–10), and (C) in red southern Chile Los Ríos and Los Lagos regions (farms 1–8). The numbers indicate the farm sample (detail in Table 1).

Table 1.

Characteristics of the Study Area Where Samples Were Collected

Categories	Animal source(s) and system type	Study area (farm number)1	Sample type (No. of sample)2	No. of sites areas	No. of phages identified on a given serovar (percentage)3				Total isolated phage	Phage isolation rate4	Phage selected
					Infantis	Heidelberg	Typhimurium	Enteritidis
Backyard cattle-livestock	Cattle in free range dual purpose (milk and meat production)	Easter Island (Orito (1), Maunga Otu’u (2), Vaihu (3), Rano Raraku (4), Puna Pau (5), Akahanga (6)	Animal (47)	6	14 (15.6%)	29 (32.2%)	13 (14.4%)	34 (37.8%)	90	(90/47)	sp1, sp2
Small and medium cattle-livestock	Cattle in free range for milk production	Central area, LGBO region Rengo (2,3), Malloa (4,5), Quinta de Tilcoco (1,6), Chimbarongo (7–10)	Animal (202), and environmental (58)	13	1 (4%)	3 (11%)	0	23 (85%)	27	(27/260)	sp7, sp8, sp11, sp23, sp24
Large farms cattle-livestock	Cattle in cage for milk production	Southern zone. Los Ríos (1,2,5,6,8) Y Los Lagos region (3,4,7)	PI	PI	3 (6,7%)	9 (20%)	4 (8,9%)	29 (64.4%)	45	(45 PI)	sp9, sp22
Total	3	3	307	19	18	41	17	86	162	162/467	9

¹ Location is described in the map (Fig. 1).

² Animal sample (bovine feces) or/and environmental (beds, floors, and drinking fountains).

³ Used by four serovars: Enteritidis (FSL S5-371), Infantis (FSL S5-506), Typhimurium (FSL S5-370), and Heidelberg (FSL S5-455). These Salmonella hosts were provided by the Food Safety Laboratory at Cornell University (Ithaca, NY) and were used for phage isolation.

⁴ No. of phage obtained/No. of sample.

FSL, Food Safety Laboratory; LGBO, Libertador General Bernardo O’Higgins; PI, previously isolated.

Sample collections

Samples were collected with Cary-Blair transport medium (COPAN, Brescia, Italy), from each animal per rectum after gentle massage of the area. Environmental samples of fecal material were collected from the holding area, manure storage, and animal bedding and deposited in sterile flasks of 60 mL using a previously described protocol (Dueñas et al.¹⁴). Each sample was immediately identified with a unique number, stored at 4°C, and transferred to the laboratory for processing within 24 h.

Salmonella

All samples were cultured for Salmonella isolation using protocol Salmonella isolate by enrichment previously described by Rivera et al.¹⁶ Briefly, samples were cultured in buffered peptone water (Beckton–Dickinson, Franklin Lakes, New Jersey) at 37°C for 24 h. One hundred microliters were then transferred into Rappaport Vassiliadis media (Beckton–Dickinson) supplemented with novobiocin (20 mg/mL), and 1 mL was transferred into tetrathionate media (Beckton–Dickinson) supplemented with iodine. These samples were incubated at 42°C for 24 h. Finally, a 100-μL aliquot of each selective enrichment broth was streaked onto an XLT-4 agar plate (Beckton–Dickinson) and incubated at 37°C for an additional 24 h. Salmonella-like strains isolated were confirmed by invA PCR Kim et al.²⁴

Isolation of VLPs

Briefly, we performed a four-stage protocol²⁰: (1) primary enrichment culture, (2) secondary culture by double agar layer, (3) purification, and (4) storage. For these, we used four Salmonella-host isolates from human-source at Food Safety Laboratory (FSL; Cornell University, Ithaca, New York), which were used for the enrichment, these were Salmonella Infantis (FSL S5-506) Salmonella Heidelberg (FSL S5-455), Salmonella Typhimurium (FSL S5-370), and Salmonella Enteritidis (FSL S5-371), because they are serovars that frequently circulate in animal production systems.¹⁶ Consecutively, only one type of lysis plaque morphology was selected (indicating the lytic effect of the phage on the host bacterium) and further purified until only one lysis-plaque morphology. Subsequently, this product was called as “VLPs” and was amplified (by repeating the plate culture), with a primary host to a concentration of at least 10⁸ PFU/mL and stored as a lysate in SM buffer (100 mM NaCl, 8 mM MgSO4, 50 mM Tris-Cl [pH 7.5]), and with 1% v/v chloroform at 4°C. The titers of each VLP were determined by spotting serial dilutions on the lawn of the host from which the phage was isolated, as previously described by Rivera et al.²⁰

Characterization of LP

A panel of 23 serovars of Salmonella, previously identified, was used to determine the LP (with two independent replicates).²⁰ The LP refers to the number of different hosts that each phage can lysate, for this study, we considered 23 different serotypes of Salmonella enterica including the most frequent serovars present in cattle production systems (Dublin, Enteritidis, and Javiana).^7–11 Each VLP-stock (5 µL of approximately 2 × 10⁵ PFU/mL) was applied by spot-test on each host culture, by overlay technique,²⁰ and was analyzed for the presence or absence of lysis. Finally, the LPs of each VPLs were analyzed by hierarchical clustering analysis and plotted through a heatmap, and dendrogram to represent the relationships of similarity between LP, given serovars found and each location (Fig. 1). The hierarchical groups were created using Ward’s binary distance method using R (2.10.0 version: R Development Core Team. Vienna, Austria).

Morphological, genomic, and taxonomy analysis of phage selections

Among the VLPs analyzed were selected following this selection criteria: at least two phages from each bovine-livestock system, with different LP and able to amplify in the same bacterial host including the four main hosts used for isolation (Enteritidis, Infantis, Heidelberg, and Typhimurium).

For genomic and taxonomic characterization, we performed analysis as previously described by Rivera et al.²⁰ Through morphological and genomic analysis we characterized the taxonomy isolated phages following the recommendations of the ICTV.²⁵ For this, the morphological visualization by transmission electron microscopy (TEM) was also conducted to check the taxonomy description based on Rivera et al.²⁶ Briefly, VLPs were precipitated with PEG8000. The phage particles were pretreated with 2% uranyl acetate and deposited onto 150–200 mesh carbon-coated Formvar film copper grids. Finally, the observation of each VLP was carried out with the Tecnai/12 model, Philips equipment, using a magnification of 50,000× to 100,000× at 85 kV. Finally, each VLP was assigned a morphology and stored as phage selected for further characterization.

The genetic analysis was to drive through whole genome sequencing of phage selected; for this, genomes were carried out by first purifying phage DNA using the phenol–chloroform protocol,²⁶ followed by ethanol precipitation. Then, the sequencing libraries were prepared using the Nextera XT preparation kit (Illumina, San Diego, California), and sequencing was conducted using the Illumina HiSeq platform at MicrobesNG (Birmingham, United Kingdom) and was annotated with RASTtk.²⁷

Comparative genomics

This analysis was performed through the alignment of nucleotide sequences by tBLASTx using VipTree version 4.0 software.²⁸ Two approaches were performed to evaluate the identity between phage sequences: (i) “diversity within phage-selected” alignment of sequences (user-uploaded genomes) and (ii) “all diversity” using reference genomes available at Virus-Host DB (https://www.genome.jp/virushostdb/).²⁸ In both cases, we selected the sequences (>60% cutoff of identity). The alignment views homologous regions between genomes detected by tBLASTx (E-value <1e-2). The parameters considered were query score, nucleotide identity (NI%), and genomic similarity scores of genomes (SG). In addition, all phage-selected sequences were analyzed for identity and plotted graphically with EasyFig.²⁹

The relative phylogenetics-diversity was performed through ViPTree-proteomic²⁸ and VICTOR web service (https://victor.dsmz.de).³⁰ To ViPTree-proteomic, the dates were normalized, and tBLASTx scores (SG; 0 ≤ SG ≤ 1) between viral genomes are calculated. A proteomic tree was generated by BIONJ based on the genomic distances (i.e., 1 − SG). Reference viral sequences and taxonomies are based on the GenomeNet/Virus-Host DB.²⁸

To obtain the analysis by VICTOR, the genome-based phylogeny and classification of prokaryotic viruses were generated using the amino acid sequences encoded by the terminase. For this, intergenomic distances were used to infer a balanced minimum evolution tree with branch support via FASTME including SPR postprocessing with D6 formula. Branch support was inferred from 100 pseudo-bootstrap replicates each. Trees were rooted at the midpoint and visualized with ggtree. Taxon boundaries at the species, genus, and family level were estimated with the OPTSIL program, the recommended clustering thresholds³⁰ and an F-value of 0.5 (fraction of links required for cluster fusion).

Results

Phages were isolated from different cattle-livestock and geographical sites. For this study, we analyzed a total of 307 samples from 19 sites/area in Chile, and of them: 47 samples from Easter Island, 260 samples from small farms in LGBO, and 45 samples previously obtained by our group and provided for this study.¹⁴ Of these, 162 Salmonella VLPs were obtained (Table 1). Geographically, the origin of each VLP was as follows: 90 VLPs from backyard cattle livestock in Easter Island, 27 VLPs from small farms in LGBO, and 45 VLPs from large farms in the southern regions of Los Rios and Los Lagos¹⁴ (Table 1 and Fig. 1). Each VLP was isolated and purified using one of the four most prevalent Salmonella serovar as host¹⁶: Salmonella Infantis (n = 18); Salmonella Heidelberg (n = 41); Salmonella Typhimurium (n = 17); and Salmonella Enteritidis (n = 86).

LP of VLP in various Salmonella serovars

We analyzed 162 Salmonella VLPs, which showed similarity in the LP, among small farms in LGBO and southern Chile (Los Rios/Los Lagos), differing from the LP of VLPs from Easter Island (Fig. 2). We found three VLPs clusters based on the LP, showing the LP clustering by isolation sites and/or type of system. The serovars most susceptible to the lytic effect were: Enteritidis (n = 113 phages/162), Dublin (n = 142/162), and Javiana (n = 144/162) (Fig. 2).

FIG. 2.

Heatmap dendrogram to represent the relationships of similarity between lytic effects of virus-like particles (VLPs) on Salmonella serovars by hierarchical groups were created using Ward’s binary distance method using R (2.10.0 version: R Development Core Team, Vienna, Austria). (A) Backyard cattle-livestock in Eastern Island, (B) large farms in Los Ríos y Los Lagos regions, and (C) small farms in LGBO region. We introduced a tertile-based color scale from dark blue (lowest presence of Salmonella serovars) to yellow (highest presence) by hierarchical group (e.g., study’s area). LGBO, Libertador General Bernardo O’Higgins.

Taxonomic and genomic analyses

Following LP characterization, at least five VLPs per source site were selected for further characterization. Then, using the selection criteria (detailed in the methodology). Through TEM nine selected phages were classified within: siphoviruses including vB_Se_EI_sp1 (sp1), vB_Se_EI_sp2 (sp2), and vB_Se_LL_sp9 (sp9). As podoviruses: vB_Se_LGBO_sp11, vB_Sh_LGBO_sp23 (sp23), vB_Sh_LGBO_sp24 (sp24), while the myoviruses identified were vB_Si_LGBO_sp7 (sp7), vB_Si_LGBO_sp8 (sp8), and vB_Se_LR_sp22 (sp22). All phages were sequenced and entered the NCBI platform under the Bioproject code PRJNA877770 (SP1, SP2, SP7, SP8, SP9, and SP11). Three additional phages were previously sequenced under the NCBI codes SP22 (MZ327261.1), SP23 (MT580117), and SP24 (MT580116) (Fig. 3).

FIG. 3.

Characterization of Salmonella bacteriophages. The figure includes transmission electron microscopy images and tables with genetic and taxonomic. The best matches were obtained by purchasing the complete nucleotide sequence of each phage from the NCBI nucleotide database and phenotypic (host range) description. (1) The taxonomy description was carried out considering the latest publication of International Committee on Taxonomy of Viruses (ICTV).²⁵ (2) Serovar abbreviations: AGO, Agona; ANA, Anatum; CHO, Choleraesuis; DUB, Dublin; ENT, Enteritidis; HEI, Heidelberg; INF, Infantis; JAV, Javiana; KEN, Kentucky; MON, Montevideo; NEW, Newport; ORA, Oranienburg; PAN, Panamá; SAI, Saintpaul; STA, Stanley; TYP, Typhimurium; 4,5,12:i:- and the hosts of each phage are marked in bold type.

The genome annotation revealed a small number of coding sequences (CDs) with known functions, including terminase, DNA metabolism, lysozyme, capsid, and tail genes among most phages. Many CDs were of unknown functionality, classified as hypothetical or phage proteins. In myoviruses, many genes coding for tRNA (n = 19 to n = 23) were observed (Fig. 3 and Fig. 4). Notably, among podoviruses, an integrase (n = 5 to n = 23) was identified (Fig. 3 and Fig. 4).

FIG. 4.

Genetic map to nucleotide sequences. The selected phage is to be compared by identity with the BLASTN tool, which is presented in grayscale for identities between 65% and 100% with EasyFig software, and the keys of each annotated CDs are shown with arrows at the bottom. CD, coding sequence.

Comparative genomics, taxonomic, and phylogenetic analysis of selected phages demonstrated a high diversity among the Salmonella phages

The first approach: “diversity within selected-phages” allowed us to identify phage sequences with 100% of identity: sp1, sp2, sp23, and sp24 (Fig. 4 and Supplementary Table S1), and all these were displayed on the genomic map (Fig. 4). These genomic differences were located on codified CDs with no described functionality (hypothetical) and tail proteins (Fig. 4). Genomic identity of sequences sp1 and sp9 showed 67.1% sp1 and sp11 a 66.2% and to sp7, sp8, and sp22 < 60% of identity (Fig. 4 and Supplementary Table S1).

To the second approach: “all diversity” used as sequence template sp1, sp22, sp23, and sp8 against sequences available in Virus-Host DB (>60% cutoff identity). Then, to sp1 and sp2, we obtained 253 matches; sp23 with 55 matches including sp24 (100%); sp22, we obtained 36 matches including sp7 (94.9%); and to sp8 nine matches.

Phylogenetic analysis was performed using two tools: VipTree-proteomic and VICTOR-Terminase. We obtained a phylogenetic relationship with 232 phage sequences available in Virus-Host DB with VipTree.²⁵ The phage sequences did not cover any viral family, but it was possible to identify a match with a host group (class-level Proteobacteria) (Fig. 5). These results were plotted on a phylogenetic tree (Fig. 5), which showed in concentric rings of different colors the taxonomic groups of phages and probable hosts (Fig. 5). While using the VICTOR tool, phylogenetic relationships were obtained between the selected phages and the viral database (Fig. 6). This tree was formed with a distance between branches of 0.2, with which we obtained: cluster I (sp1 and sp2) formed by sequences of the siphovirus morphotype obtained from the same geographical and productive origin, and were 100% related at the level of family, genus, and viral species (Fig. 6). Cluster II (sp23 and sp24) grouped podoviruses from the same origin and were also 100% related while cluster III included sequences (sp7 and sp22) <95% related (Fig. 6). Finally, the combined analysis of the results of VLPs origin, LP, taxonomy, and phylogeny did not allow to obtain a classification that could be extrapolated to all cases.

FIG. 5.

Overall phylogenetic tree. The diversity of the phage selected, with respect to 232 phage sequences available in the virus-host DB dsDNA database on ViPTree software. The origin of each clade of the tree is marked with different colors, and each phage selected in this study is marked with red stars.

FIG. 6.

(A) Phylogenetic characterization of selected phages for terminase was obtained with the VICTOR web service tool (https://victor.dsmz.de). In this case, three clusters were obtained (I, II, and III), and the distance is marked with a black bar (0.2). The level of relationship between the branches of the tree is shown in different colors marked in the image as: Family (F), Subfamily (Sf), Genus (G), and Species (Sp), the origin of the phages and morphotype was shown with the symbol of phages siphovirus, myovirus, and podovirus and the same colors used in the map: green (Easter Island) blue LGBO and red (Southern Chile). (B) Lysis profile obtained by dendrogram (Fig. 2) considering serovars: AGO, Agona; ANA, Anatum; CHO, Choleraesuis; DUB, Dublin; ENT, Enteritidis; HEI, Heidelberg; INF, infantis; JAV, Javiana; KEN, Kentucky; LGBO, Libertador General Bernardo O’Higgins; MON, Montevideo; NEW, Newport; ORA, Oranienburg; PAN, Panamá; SAI, Saintpaul; STA, Stanley; TYP, Typhimurium; 4,5,12:i:-.

Discussion

It is noteworthy that from none of the 307 samples analyzed in this study we isolated Salmonella using the enrichment method that has been successful in isolating Salmonella strains from other animal systems in Chile,¹⁶ and this same finding was previously described by our research group in dairy cattle in southern Chile.¹⁴ In addition, there is evidence to support the hypothesis of Salmonella circulation as a postmortem finding in dairy cattle in southern Chile.³¹ In this sense, one of the difficulties of Salmonella isolation in cattle is that after infection, it can be retained in lymph nodes.^3,4 Thus, the present study investigates the presence of Salmonella phages in the absence of Salmonella strains in three different cattle systems. For this, 162 VLPs were characterized to understand phenotypic LPs diversity in several Salmonella serovars likely circulating in these sites. It was found that a high percentage of VLPs exhibited an LP against Enteritidis (70%), Dublin (88%), and Javiana (88%), regardless of the study sites (Fig. 2). In this sense, Salmonella Dublin is one of the most frequently described in other countries.^8–11 In addition, LP analysis through heatmap showed that LP of VLPs from Easter Island’s had the widest LP compared with those obtained from other production systems (Fig. 2). Subsequently, it was explored whether the phenotypic diversity described for the LPs was related to a diversity among the selected phage sequences and relatively compared to the Salmonella phage database. In this regard, we acknowledge the limitations of our study in terms of the low number of phages sequenced (9 phages/162 VLPs) and the choice of using the classical method of phage isolation and characterization versus metagenomics.¹² Finally, our approach allows us to integrate metadata information from each study site with the phenotypic and genetic characteristics of each phage, which in the future could help generate models to predict the presence of specific Salmonella serovars in each system.

The characterization of taxonomic is challenging, mainly due to the low percentage of gene sequences with recognized functionality (approximately 15–20% of the complete genome).¹² Therefore, many approaches were used in this study to describe the taxonomy and diversity of selected phages.²¹ The taxonomic analysis (by TEM and comparative genomic) showed three morphotypes: (i) siphoviruses (n = 3); (ii) myoviruses (n = 3); and (iii) podoviruses (n = 3) and demonstrated that our phages imply a relative diversity (into all Salmonella phages available).

The comparative genomics revealed diversity of Salmonella phages, in terms of NI and phylogeny analyses. Both analyses showed that sequences with the same phylogenetic origin also had NI (>60% identity in all cases), specifically in sp1–sp2 and sp23–sp24. In contrast, the sp8 showed lower identity with selected sequence and with other phages from date bases. In addition, this sequence did not cluster within any group in the phylogenetic analysis.

In particular, the comparison between siphoviruses (sp1 and sp2) showed only one difference regarding their tail protein, which could be probably linked to the differing host range.^19,21 These phages also showed the best identity with the Salmonella phage vB_Se_STGO-35–1 (NC_054648.1), previously isolated in Chile, from backyard poultry production systems.^20,26 While other siphovirus sp9 showed an identity of 67.1% with phage sp1 and phage sp2, and sp9 showed an identity of 97.4% with Salmonella phage PSDA-2 (MK214385.1)³² (Fig. 3). In addition, these siphoviruses have similar genetic sizes with high genetic identity but different LPs (Fig. 3 and Supplementary Table S1).

The podoviruses sp23 and sp24 showed high overall identity (100%) and interestingly, we found integrases in podoviruses, with transduction capacity that have been previously noted for some of the representatives of this family³³ (Fig. 3). However, no significant identity was found between sp11 and sp23. The myoviruses sp22 and sp7 presented a genome size consistent with typical Salmonella phage myoviruses³⁴ and best identity matched with Ounavirinae Salmonella virus VSe102 (MG251392.1), while sp8 presented a smaller genome size which is genetically related to another phage JEP7 (MT764207.1),³⁵ and best matched with genus Rosemountvirus of the family myoviridae, isolated from a soil sample near a poultry farm.

The phylogenetic analysis showed a close relationship between siphovirus sp1 and sp2 phages (cluster I), at family, genus, and species levels (Fig. 6), both phages were isolated from the same origin and showed 100% identity with small genetic differences (Fig. 4). The other siphovirus (sp9) presented an identity of 97.1%, however, phylogenetically it was not related, which was demonstrated in the phylogenetic analysis in Figure 5 and Figure 6, which can be explained by the fact that the identity analysis was performed with the total nucleotide sequences of phages, and the phylogenetic analysis was directed to the coding of probable protein sequences, so this last analysis has a greater predictive value to explain the diversity of the selected phages.^12,18,19,21 Therefore, it is quite interesting to highlight that although these phages are identical in terms of nucleotide comparison, they can present phenotypic and phylogenetic differences that confer different behavior with the host.^18,19,21

In the case, podovirus sp23 and sp24 also had a 100% identity and a close genetic relationship in both analyses (Fig. 5 and Fig. 6) and with respect to the other podovirus sp11, it did not present identity with other podoviruses (Supplementary Table S1) but had with the phages siphovirus sp1 and sp2, with respect to its relative genetic diversity it formed the same cluster II (Fig. 6), in the phylogeny analysis with the terminase.

Finally, for myovirus phages, the identity comparison between them (sp22 and sp7) was 94.9% (Supplementary Table S1); however, again phylogenetic analyses indicate that this similarity is relative using protein prediction, in the case of whole genome analysis their prediction was divergent (Fig. 5) and in the terminase analysis it was included in the same cluster (Cluster III) (Fig. 6).

All these comparative genomic analyses were not conclusive to characterize the viral family consistent with the postulates of the ICTV.²⁵ However, it was possible to classify the selected phages by morphotype and class (Caudoviricetes).³⁰ The clusters I and II, obtained in Figure 6, showed the same origin of isolation, genetic relationship, and similar LP, but this observation could not be generalized to all the phages analyzed, so in overall terms, there was no relationship between genotype, livestock type, or geographic location. This could occur if different environmental sites shared similar microbial isolates but their geographic separation or different niches facilitated local coevolution to occur, which is visualized by phenotypic divergence below the taxonomic classification of viral species (e.g., difference in LP).^19,21 In general, the nesting and modular conformation of phage genomes are not mutually exclusive; thus, a cluster could contain phages with LP phenotypic differences.²¹

Conclusions

We conducted this study to understand the presence of Salmonella phages in the absence of Salmonella strains using phenotypic and genomic methodologies. The phenotypic analysis, LP, indicated differences across cattle production systems, showing main susceptibility to the Salmonella Dublin, Salmonella Enteritidis, and Salmonella Javiana serovars. While comparative genomics demonstrated that relying on a single methodology makes it difficult to identify phage taxonomic and diversity. However, the combined analysis of multiple methodologies provides a better understanding of the relative diversity of phage sequences, which is evident in the different groups of Salmonella phages.

The identification and characterization of Salmonella phages, along with the observed absence of Salmonella strains in three different cattle livestock environments, lay the foundation for improved detection and surveillance strategies in livestock settings. Sequencing additional phages could further enable the development of predictive and surveillance models for Salmonella presence in cattle livestock systems.

Authors’ Contributions

D.R. and A.I.M.-S.: Conceptualization and validation. D.R., F.D., and A.I.M.-S.: Funding acquisition, writing—review and editing. Methodology and writing—original draft: All authors. All authors have read and agreed to the published version of the article.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

The authors acknowledge the funding sources: (1) Agencia Nacional de Investigación y Desarrollo Fondo de Fomento al Desarrollo Científico y Tecnológico (ANID FONDECYT VIU22P0058 to D.R.), (2) (ANID FONDECYT 1231082 to A.I.M.-S.), and (3) (ANID FONDECYT 1191747 to C.H.W.).