Yersinia enterocolitica-Specific Infection by Bacteriophages TG1 and ϕR1-RT Is Dependent on Temperature-Regulated Expression of the Phage Host Receptor OmpF

ABSTRACT

Bacteriophages present huge potential both as a resource for developing novel tools for bacterial diagnostics and for use in phage therapy. This potential is also valid for bacteriophages specific for Yersinia enterocolitica. To increase our knowledge of Y. enterocolitica-specific phages, we characterized two novel yersiniophages. The genomes of the bacteriophages vB_YenM_TG1 (TG1) and vB_YenM_ϕR1-RT (ϕR1-RT), isolated from pig manure in Canada and from sewage in Finland, consist of linear double-stranded DNA of 162,101 and 168,809 bp, respectively. Their genomes comprise 262 putative coding sequences and 4 tRNA genes and share 91% overall nucleotide identity. Based on phylogenetic analyses of their whole-genome sequences and large terminase subunit protein sequences, a genus named Tg1virus within the family Myoviridae is proposed, with TG1 and ϕR1-RT (R1RT in the ICTV database) as member species. These bacteriophages exhibit a host range restricted to Y. enterocolitica and display lytic activity against the epidemiologically significant serotypes O:3, O:5,27, and O:9 at and below 25°C. Adsorption analyses of lipopolysaccharide (LPS) and OmpF mutants demonstrate that these phages use both the LPS inner core heptosyl residues and the outer membrane protein OmpF as phage receptors. Based on RNA sequencing and quantitative proteomics, we also demonstrate that temperature-dependent infection is due to strong repression of OmpF at 37°C. In addition, ϕR1-RT was shown to be able to enter into a pseudolysogenic state. Together, this work provides further insight into phage-host cell interactions by highlighting the importance of understanding underlying factors which may affect the abundance of phage host receptors on the cell surface.

IMPORTANCE Only a small number of bacteriophages infecting Y. enterocolitica, the predominant causative agent of yersiniosis, have been previously described. Here, two newly isolated Y. enterocolitica phages were studied in detail, with the aim of elucidating the host cell receptors required for infection. Our research further expands the repertoire of phages available for consideration as potential antimicrobial agents or as diagnostic tools for this important bacterial pathogen.

INTRODUCTION

Yersinia enterocolitica, a facultative anaerobic, Gram-negative, nonsporulating, short bacillus isolated frequently from soil, water, animals, and foods, is an important zoonotic pathogen leading to human and animal enteric infection (1). The main animal reservoir for Y. enterocolitica is pigs, and pork-derived products are thought to be the main source of human infections, in addition to the drinking of contaminated water and blood transfusions (1, 2). Symptoms of yersiniosis may include diarrhea, terminal ileitis, mesenteric lymphadenitis, and septicemia (3). Among the species within the genus Yersinia, Y. enterocolitica is highly heterogeneous and is grouped into six phylogroups (4). The widely used bioserotype groups form the basis of the phylogroups such that phylogroup 1 contains the biotype 1A strains, phylogroup 2 the highly pathogenic biotype 1B strains, phylogroup 3 the bioserotype 4/O:3 strains, phylogroup 4 the bioserotype 3/O:9 strains, phylogroup 5 the bioserotype 2/O:5,27 strains, and phylogroup 6 the serotype O:2,3 strains that are rarely isolated from hares (4–7). Y. enterocolitica is also represented by over 60 serotypes that are determined by the variability of O antigens present in the outer cell membrane (8, 9). The predominant pathogenic strains associated with yersiniosis belong to bioserotypes 1B/O:8, 2/O:5,27, 2/O:9, 3/O:3, and 4/O:3, with the last being the most common in Europe, Japan, Canada, and the United States (1, 2). From 2010 to 2012, 98% of all reported yersiniosis infections worldwide were acquired in Europe, and most (97%) were caused by Y. enterocolitica, with the remainder caused by Yersinia pseudotuberculosis (10). In 2015, the most commonly reported Y. enterocolitica serotype in the European Union was O:3 (89%), followed by serotypes O:9 (7%), O:5,27 (2%), and O:8 (2%) (10).

Although several bacteriophages infecting Y. enterocolitica have been described, few have been studied in detail to provide reliable information on morphology, host range, and/or receptor specificity. To date, bacteriophages ϕYeO3-12 (11–13) and vB_YenP_AP5 (14) with specificity for Y. enterocolitica O:3, phage PY54 exhibiting a host range restricted to Y. enterocolitica O:5 and O:5,27 (15), Yersinia phage ϕR1-37 with a broad host range within the species Y. enterocolitica (16, 17), and Yersinia phage PY100 (18) exhibiting a broader host range restricted to the genus Yersinia have been described. These bacteriophages use different parts of the Y. enterocolitica lipopolysaccharide (LPS) as receptors (19). Analysis of the host range combined with genetic and structural data has shown that the receptor for ϕR1-37 is the Y. enterocolitica O:3 LPS outer core (OC) hexasaccharide (16). The host receptor for phages ϕYeO3-12 and vB_YenP_AP5 has been determined to be the LPS O antigen of serotype O:3, consisting of the sugar 6-deoxy-l-altropyranose (12, 14, 20). Given the interest in bacteriophages because of their potential use as therapeutic, diagnostic, and biocontrol agents, the aim of this study was to characterize two newly isolated bacteriophages that are active against several epidemiologically significant Y. enterocolitica serotypes. In this study, the genome characterization, morphology, host range, host cell receptor specificity, and taxonomic position of the myovirus phages vB_YenM_TG1 (here called TG1) and vB_YenM_ϕR1-RT (here called ϕR1-RT) are described.

MATERIALS AND METHODS

Bacterial strains, phage isolation, and growth conditions.

Bacterial strains, bacteriophages, and plasmids are listed in Table 1. Bacteriophage ϕR1-RT was isolated from the incoming sewage of the Turku (Finland) city sewage treatment plant, as described for other viruses (19), whereas bacteriophage TG1 was isolated from pig manure collected from a rural farm in Ontario, Canada, as described previously for the isolation of Y. enterocolitica phages for phage typing (21). For DNA extraction and morphological studies, ϕR1-RT was propagated on Y. enterocolitica strain YeO3-R1 (22) and TG1 was propagated on Y. enterocolitica strain YeO3-c (23).

TABLE 1.

Bacterial strains, plasmids, and bacteriophages

Strain, plasmid, or bacteriophage	Comments	Reference or source
Strains
Y. enterocolitica
6471/76 (YeO3)	Serotype O:3, wild type; human stool isolate	23
6471/76-c (YeO3-c)	Virulence plasmid-cured derivative of YeO3	23
YeO3-ϕR1-RT-R2	ϕR1-RT-resistant spontaneous derivative of YeO3	This work
YeO3-ϕR1-RT-R7	ϕR1-RT-resistant spontaneous derivative of YeO3	This work
YeO3-ϕR1-RT-R9	ϕR1-RT-resistant spontaneous derivative of YeO3	This work
YeO3-R1	(=YeO3-c-R1) Spontaneous rough strain	22
YeO3-hfq::Km	Hfq::Km-GenBlock, Kmr; urease negative	Leskinen et al., submitted for publication
YeO3-R1-Cat17	ompF::Cat-Mu derivative of YeO3-R1	This work
YeO3-R1-Cat17::pSW25T_OmpF	cis-complemented ompF::Cat-Mu strain	This work
YeO3-R1-M164	waaF::Cat-Mu derivative of YeO3-R1, Clmr	79
YeO3-R1-M196	galU::Cat-Mu derivative of YeO3-R1, Clmr	79
YeO3-R1-M205	hldE::Cat-Mu derivative of YeO3-R1, Clmr	79
YeO3-c-OC		25
YeO3-c-OCR		25
K14	Serotype O:9
gc815–73	Serotype O:5,27	80
8081	Serotype O:8	81
Escherichia coli
BL21 Star (DE3) PLysS		Invitrogen
DH10B
Plasmids
pTM100		82
pTM100_OmpF	Complementation plasmid with wild-type ompF gene cloned into pTM100	This work
pSW25T	Suicide vector	83
pSW25T_OmpF	Complementation suicide plasmid with wild-type ompF gene cloned into pSW25T	This work
pCDF Duet-1	pCloDF13 replicon, T7lac promoter and 2 MCSs, each with an optional N-terminal His6 tag sequence; streptomycin resistance marker	Novagen
pET21a(+)	ColE1 (pBR322) replicon, T7lac promoter, N-terminal T7 tag sequence, and optional C-terminal His6 tag sequence; ampicillin resistance marker	Novagen
pCDF Duet-1 Gp37	Phage TG1 ORF250 (bp 4 to 1830) cloned in frame into MCS1 of pCDF Duet-1 for expression of N-terminal His6-tagged protein Gp37	This study
pCDF Duet-1 Gp37-Gp38	Phage TG1 ORF251 cloned into MCS2 of pCDF Duet-1 Gp37 for coexpression of N-terminal His6-tagged Gp37 and tail fiber assembly chaperone Gp38	This study
pET21a(+) Gp57A	Phage TG1 ORF143 cloned into MCS of pET21a(+) for expression of general trimerization chaperone Gp57A	This study
Bacteriophages
TG1		This study
ϕR1-RT		This study

Electron microscopy.

The preparation of the phage particles for transmission electron microscopy (TEM) was done as described previously (17, 24). Details are presented in the supplemental Materials and Methods.

Host range.

The lytic activities of ϕR1-RT and TG1 were tested on 109 and 160 strains (see Table S1 in the supplemental material), respectively, belonging to 13 Yersinia species, as determined by standard spot tests (24). Briefly, 10 μl from a phage suspension containing approximately 10⁸ PFU was spotted in the middle of a lawn of bacteria and incubated for 18 to 24 h. Each strain was tested three times at 25°C and at 37°C. Bacterial strains were considered sensitive to the phage if the degree of lysis was observed as a complete clearing, clearing throughout but with a faint hazy background, substantial turbidity throughout the cleared zone, or a few individual plaques (24). Bacterial strains were considered resistant if there was no effect of the phage on bacterial growth.

Genome sequencing and assembly.

Details of the determination of the genomic sequences of phages ϕR1-RT and TG1 as well as the draft genomes of Y. enterocolitica strains YeO3-ϕR1-RT-R2, -R7, and -R9 are presented in the supplemental Materials and Methods.

Bioinformatics.

A detailed description of the bioinformatics tools used is given in the supplemental Materials and Methods.

Complementation of the Y. enterocolitica O:3 OmpF mutant.

The full open reading frame (ORF) of the ompF gene plus the upstream promoter region of YeO3-c was cloned as a 2-kb PCR fragment that was amplified with Phusion DNA polymerase using primer pair OmpC-F2 and OmpC-R2 (see Table S2 in the supplemental material) into plasmids pTM100 and pSW25T to obtain plasmids pTM100_OmpF and pSW25T_OmpF, respectively (Table 1). Briefly, the PCR fragments were digested with MfeI and ligated with EcoRI-digested, shrimp alkaline phosphatase-treated pTM100 or pSW25T. The constructed plasmids were mobilized to the OmpF mutant strain YeO3-c-R1-Cat17 by diparental conjugation as described earlier (25).

Phage adsorption assay.

To identify the phage cell host receptors, a variety of Y. enterocolitica O:3 mutants (Table 1) were utilized in phage adsorption experiments. Approximately 5 × 10³ PFU of phage ϕR1-RT or phage TG1 in 100 μl was mixed with a 400-μl sample of bacteria (A₆₀₀, ∼1.2). The suspension was incubated at room temperature for 5 min and centrifuged at 16,000 × g for 3 min, and the phage titer remaining in the supernatant, i.e., the residual PFU percentage, was determined. LB was used as a nonadsorbing control in each assay, and the phage titer in the control supernatant was set to 100%. Each assay was performed in duplicate and repeated at least three times.

Total RNA extraction and RNA sequencing.

A detailed description of the RNA extraction and sequencing methods is presented in the supplemental Materials and Methods. The RNA sequence data have been deposited in the Gene Expression Omnibus (accession number GSE66516).

Quantitative proteomics.

A detailed description of the quantitative proteomics methods is presented in the supplemental Materials and Methods.

Transduction assay.

Y. enterocolitica O:3 strain YeO3-hfq::Km with the hfq gene knocked out with a kanamycin resistance cassette (Table 1) was used as a donor, and transducing particles were produced by infecting this strain with phage ϕR1-RT by use of the soft agar overlay method. Following overnight incubation, phages were eluted from the soft agar using SM buffer (5.8 g of NaCl per liter, 2.0 g of MgSO₄·7H₂O per liter, 50 mM Tris-HCl [pH 7.5]). The transducing lysates were centrifuged and treated with chloroform to prevent contamination with the donor strain. The titer of the obtained transducing stock was 6.62 × 10⁹ PFU/ml. Y. enterocolitica strain 6471/76 was used as the recipient. For the transduction of the recipient strain, 10 1-ml aliquots of log-phase bacterial cultures containing 10⁹ CFU/ml cells were mixed with 100 μl of 10⁻² diluted transducing phage stock, with a resulting multiplicity of infection (MOI) of 0.006. After 15 min, the bacterial cells were centrifuged and washed with LB and then centrifuged to remove the unabsorbed phages. The final cell pellet was resuspended in 100 μl LB, and the cells were allowed to recover during 30 min of incubation with vigorous shaking. Subsequently, the bacterial cultures were plated on urea agar plates (0.1% peptone, 0.1% glucose, 0.5% NaCl, 0.2% KH₂PO₄, 0.00012% phenol red, 2% urea, 1.5% agar) supplemented with kanamycin (200 μg/ml) and incubated for 48 h. The kanamycin-resistant and urease-negative colonies were considered transduced. The transducing stock was also plated to ensure no contamination with the donor strain.

Growth curves.

Overnight bacterial cultures were diluted 1:10 in fresh LB medium, and 180-μl aliquots were distributed into honeycomb plate wells (Growth Curves Ab Ltd.), where they were mixed with 20-μl aliquots of different ϕR1-RT phage stock dilutions (10⁰ to 10⁻⁴). A negative control was obtained by mixing 20 μl of phage stock with 180 μl of medium, whereas the positive control consisted of 180 μl of bacterial culture and 20 μl of medium. The growth experiments were carried out at 4°C, 10°C, 16°C, 22°C, and 37°C using a Bioscreen C incubator (Growth Curves Ab Ltd.) with continuous shaking. The optical density at 600 nm (OD₆₀₀) of the cultures was measured at selected time intervals. The averages were calculated from values obtained for the bacteria grown in five parallel wells.

Isolation of phage-resistant mutants.

A culture of wild-type Y. enterocolitica strain 6471/76 was used to flood LB agar plates (LA). After the excess fluid was removed, the plates were allowed to dry before two 100-μl aliquots of the ϕR1-RT stock were pipetted on the lawn of cells. The plates were incubated at 22°C and inspected daily for phage-resistant colonies growing within the lysis zones. After 3 days, several colonies appeared, and among them, three confirmed phage-resistant derivatives were isolated. The strains were named YeO3-ϕR1-RT-R2, YeO3-ϕR1-RT-R7, and YeO3-ϕR1-RT-R9.

Cat-Mu library screening.

The Cat-Mu transposon insertion library in Y. enterocolitica strain YeO3-R1 has been described previously (26, 27). In the present work, a library representing 16,000 independent insertion mutants was screened. The library was grown in LA supplemented with 100 μg/ml chloramphenicol (LA-Clm) until an OD₆₀₀ of ∼0.5 was reached. Phage ϕR1-RT was added to 1 ml of the library culture at an MOI of ∼10, fresh LB was added to achieve 5 ml, and the culture was incubated at 22°C for 2 h, during which time all phage-sensitive bacteria were expected to be infected and lysed. The surviving bacteria were pelleted by centrifugation, washed twice with 1 ml LB to remove the remaining phages, and after resuspension into 100 μl of LB, plated on four LA-Clm plates that were incubated at 22°C. The Clm^r colonies were restreaked on LA-Clm plates for further study.

Arbitrary PCR.

A detailed description of the arbitrary PCR method is presented in the supplemental Materials and Methods.

Cloning, expression, and purification of the phage long-tail fiber host RBP.

The phage TG1 distal long tail fiber (LTF) protein Gp37 was coexpressed with phage-encoded chaperones Gp57A and Gp38 to synthesize the native form of the putative receptor binding protein (RBP) as described previously for the LTF of phage T4 (28). The Gp37-encoding gene was first cloned into multiple cloning site (MCS) 1 of pCDF Duet-1 (conferring streptomycin resistance), producing pCDF Duet-1 Gp37. Then, the Gp38-encoding gene was cloned into the MCS 2 of pCDF Duet-1 Gp37, yielding pCDF Duet-1 Gp37-Gp38. Plasmid pET21a(+) conferring ampicillin resistance was used to clone the chaperone Gp57A-encoding gene, yielding plasmid pET21a(+) Gp57A. The plasmid constructs carry, under the control of promoter T7, high-level inducible gene expression with a His₆ fusion tag at the N terminus for purification by chelating affinity chromatography (see Fig. S1 in the supplemental material). The genes encoding Gp38 and Gp57A, however, were expressed without a purification tag. PCR, restriction analysis, and DNA sequencing were used to verify the structure of the plasmids. For expression, Escherichia coli BL21 Star (DE3) pLysS cells (Invitrogen) were transformed with pCDF Duet-1 Gp37 or pCDF Duet-1 Gp37-Gp38, and the same plasmids were also cotransformed with pET21a(+) Gp57A. Plasmid-bearing E. coli was grown aerobically at 37°C to an OD₆₀₀ of ∼0.6 with shaking at 200 rpm in 250 ml of 2× YT medium (16 g/liter tryptone, 10 g/liter yeast extract, 5.0 g/liter NaCl, 0.22-μm filter sterilized, pH 6.5 to 7.5) supplemented with 50 μg/ml of ampicillin and/or 50 μg/ml streptomycin as required. Protein expression was induced by the addition of 1 mM isopropyl-d-1-thiogalactopyranoside (IPTG) (Sigma-Aldrich, USA) and incubation for 24 h at 30°C with shaking at 200 rpm. Cells were harvested by centrifugation at 10,000 × g for 15 min at 4°C, and the pellets were resuspended in 25 ml of buffer A (50 mM sodium phosphate, 300 mM NaCl, 10 mM imidazole, pH 8.0) supplemented with a protease inhibitor cocktail (Roche). Cells were disrupted by 10 rounds of 15 s of sonication using a Virsonic Digital 475 ultrasonicator (VirTis, NY, USA), alternating with incubation on ice. Insoluble debris was removed by centrifugation at 18,000 × g for 30 min at 4°C, and the soluble fraction was filtered through a 0.22-μm-pore-size filter (EMD Millipore, USA). The protein was purified by immobilized metal ion affinity chromatography using a nickel-nitrilotriacetic acid (Ni-NTA) agarose column (Novex, Invitrogen) according to the manufacturer's protocol. Captured proteins were eluted from the column using buffer B (50 mM sodium phosphate, 300 mM NaCl, 500 mM imidazole, pH 8.0) and concentrated using Amicon-Pro centrifuge filters (Millipore) with a 10,000-Da molecular mass exclusion limit and incorporating three washes with 10 mM Tris-HCl (pH 8.5). Protein concentration was estimated by measuring sample absorbance at 280 and 260 nm using a Nanodrop 2000 UV-Vis spectrophotometer (Thermo Scientific, USA) and a Qubit protein assay kit with a Qubit 1.0 fluorometer (Life Technologies) per the manufacturer's instructions. Protein analysis was performed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (29) using Mini-Protean TGX stain-free precast gels (Bio-Rad Laboratories, USA) and Coomassie blue staining. Precision Plus Protein unstained standard (Bio-Rad Laboratories, Inc., Hercules, CA, USA) was used as a size marker for the molecular analysis of proteins. Analysis of protein bands and molecular weight (MW) estimates was performed using a Molecular Imager Gel Doc XR+ system (Bio-Rad Laboratories, Inc., Hercules, CA, USA) and Quantity One software (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Accurate MW determinations and peptide mass fingerprinting analysis were performed via mass spectrometry (MS) at the Mass Spectrometry Facility, Advanced Analysis Centre of the University of Guelph (Ontario, Canada).

Cell decoration with bacteriophage host recognition binding proteins.

Confocal laser immunofluorescence microscopy was used to visualize the binding of the phage TG1 LTF protein Gp37 to Y. enterocolitica by following methodology described by others (30). Yersinia strains grown in Trypticase soy broth (TSB) at 25°C or 37°C for 24 h were resuspended in wash buffer (50 mM Tris-HCl, pH 7.5), and 10-μl volumes were spotted onto clean glass slides. After air drying, the cells were fixed in a solution of 5% glutaraldehyde for 10 min and blocked with blocking buffer (5% bovine serum albumin [BSA] in 50 mM Tris-HCl buffer, pH 7.5) for 10 min. The slides were then incubated for 1 h in a solution containing 10 μg/ml of phage TG1 Gp37 (prepared in blocking buffer), followed by washing three times for 5 min in wash buffer. The slides were then incubated for 1 h in anti-His₆ tag (HIS.H8) mouse monoclonal antibody solution prepared in blocking buffer (1:1,000 dilution) and washed three times for 5 min with wash buffer. In a dark room, the slides were then incubated for 1 h in goat anti-mouse IgG DyLight 488 polyclonal antibody solution (1:500) prepared in blocking buffer and washed three times for 5 min with wash buffer. The slides were air dried prior to analysis. Cells were imaged using an upright Leica DM 6000B confocal laser microscope connected to a Leica TCS SP5 system. Images were collected digitally using Leica LAS AF imaging software and processed using ImageJ (31). To verify the specificity of the fluorescent signal, control samples were immunolabeled as described above, with the omission of incubation with the primary antibody. All antibodies were acquired from Pierce Scientific, USA.

Accession number(s).

The complete genome sequences of Yersinia phages vB_YenM_TG1 and vB_YenM_ϕR1-RT were deposited in the NCBI nucleotide database (GenBank) under accession numbers KP202158 and HE956709, respectively. The RNA sequence data have been deposited in the Gene Expression Omnibus database under accession number GSE66516.

RESULTS

Phage morphology.

Phages ϕR1-RT and TG1 were negatively stained and examined by TEM. Both phages exhibited a prolate head with apparent icosahedral symmetry and a tubular contractile and rigid tail showing transverse striations (Fig. 1). The average dimensions for the ϕR1-RT head are 82 ± 4 nm short edge-to-edge and 101 ± 5 nm vertex-to-vertex, and the tail, including the baseplate, is on average 130 ± 7 nm long. The average dimensions for the TG1 head are 91 ± 2 nm short edge-to-edge and 115 ± 6 nm vertex-to-vertex, and the tail, including the baseplate, is on average 129 ± 1 nm long. Collectively, these morphological features indicate that these phages belong to the Myoviridae family.

FIG 1.

Electron microscopy images showing bacteriophage ϕR1-RT and TG1 morphologies. (A) ϕR1-RT virions at ×39,440 magnification. The virion head and tail are indicated, as well as long tail fibers (LTF) and a baseplate with protruding tail pins (labeled B). Scale bar, 100 nm. (B) A ϕR1-RT virion at ×84,320 magnification. A baseplate with protruding tail pins (labeled B) and a neck and collar with neck fibers (labeled A) can be observed. Scale bar, 50 nm. (C) A ϕR1-RT virion at ×108,800 magnification. Suggested LTF can be seen bent up toward the head along the tail sheath, as described for bacteriophage T4 (84). Scale bar, 50 nm. (D) Bacteriophage TG1 virion shown at ×150,000 magnification. A neck and collar with neck fibers (labeled A), a baseplate with protruding tail pins (labeled B), and an extended LTF can be observed. The scale bar indicates size in nanometers.

Host specificity.

The host range of phages TG1 and ϕR1-RT were determined by testing their lytic activity on 160 and 109 strains, respectively, belonging to 13 Yersinia species, revealing virulence for Y. enterocolitica strains of serotypes O:1, O:2, O:3, O:5, O:6, O:5,27, O:7,8, and O:9 and some strains of serotype O:6,30 and O:6,31, while strains from other Y. enterocolitica serotypes and species within the genus Yersinia were resistant to phage infection (Table 2). TG1 and ϕR1-RT lysed their host when grown at 25°C but not at 37°C. Additionally, TG1 was unable to infect strains belonging to other 20 other genera (see Table S3 in the supplemental material), demonstrating that the phages' host range is restricted to Y. enterocolitica.

TABLE 2.

Lytic activities of phages TG1 and ϕR1-RT^a

Yersinia species	Phage-sensitive serotypesb	Serotypes with phage-sensitive (S) and -resistant (R) strains	Phage-resistant serotypesc
Y. enterocolitica	O:1 [2], O:2 [2], O:3 [16], O:5 [9], O5,27 [10], O:6 [2], O:7,8 [2], O:9 [13]	O:6,30 [1S/2R], O:6,31 [1S/1R]	O:1,2,3 [1], O:4 [1], O:4,32 [1], O:8 [14], O:10 [4], O:13 [1], O:13a,13b [1], O:13,7 [2], O13,18 [1], O:14 [1], O:20 [2], O:21 [3], O:25 [1], O:25,26,44 [1], O:26,44 [1], O28,50 [1], O:34 [1], O:35,36 [1], O35,52 [1], O:41(27),K1 [1], O41(27),42 [1], O:41(27),42,K1 [1], O:41,43 [1], O:41(27),43 [2], O:50 [1], K1 NT [2], NT [3]
Y. aleksiciae			O:16 [2]
Y. aldovae			UT [2]
Y. bercovieri			O:58,16 [2], NT [1], UT [2]
Y. frederiksenii			O:3 [1], O:16 [1], O:35 [1], O:48 [1], K1 NT [1], NT [1], UT [2]
Y. intermedia			O16,21 [1], O:52,54 [1], UT [2]
Y. kristensenii			O:3 [1], O:12,25 [1], NT [2], UT [3]
Y. mollareti			O:3 [1], O:59(20,36,7) [1], UT [2]
Y. nurmii			UT [1]
Y. pekkanenii			UT [1]
Y. pseudotuberculosis			I [2], O:1b [2], O:3 [2]
Y. rohdei			UT [2]
Y. ruckeri			NT [1], UT [5]

^aSensitivity was tested for 160 Yersinia species strains (see Table S1 in the supplemental material) at 25°C. Phage ϕR1-RT sensitivity was tested with only the 109 University of Helsinki (UH) source strains (see Table S1).

^bThe numbers of strains studied are given in brackets.

^cNT, nontypeable and either cross-reacting or not agglutinating with Y. enterocolitica O:3, O:5, O:8, or O:9 antisera; UT, untyped.

General features of the phage genomes.

The genome of phage ϕR1-RT is 168,809 bp long and has a GC content of 34.5%. The genome contains 262 ORFs, of which 217 genes are carried on the reverse strand (as displayed on the genetic map) and 45 genes on the forward strand (see Fig. S2 in the supplemental material), with sizes ranging from 117 bp (with a product of 38 amino acids) to 3,738 bp (with a product of 1,245 amino acids). The genome of TG1 is smaller than that of ϕR1-RT at 162,101 bp in length but with a similarly low GC content of 34.6%. The TG1 genome also contains 262 ORFs, of which 223 genes are transcribed on the reverse strand (as displayed on the genetic map) and 39 genes on the forward strand (see Fig. S3), with sizes ranging from 114 bp (with a product of 37 amino acids) to 3,099 bp (with a product of 1032 amino acids). The GC content of these phages is significantly lower than that associated with the host, which has a GC content ranging from 47.1% ± 0.2% (32) to 48.5% ± 1.5% (33). The two genomes contain the same additional four tRNA genes (Gly_GGA, Trp_TGG, Arg_AGA, Met_ATG) identified using tRNAScan (34) and ARAGORN (35). Constitutively low-GC phage genomes are often supplemented with tRNA genes that, once expressed, enhance translation efficiency when infecting high-GC-content hosts (36). At the DNA level, the TG1 genome shows 98% identity with a query coverage of 93%, for an overall DNA sequence identity of 91% with ϕR1-RT. All ORFs were screened using the BLASTP and PSI-BLAST algorithms (37, 38). Based on protein homology, putative functions could be assigned to 121 (46%) gene products of phage TG1 and 115 gene products (44%) of phage ϕR1-RT. Most of the identified homologs are conserved among T4-like phages and are either structural or involved in DNA replication, recombination, repair, or packaging. Thus, the phage T4 gene nomenclature was used to name these genes (see Table S4 in the supplemental material).

DNA replication, recombination, and repair.

Numerous genes that play a direct role in DNA replication, recombination, and repair were identified within the phage ϕR1-RT and TG1 genomes. Among the genes directly involved in DNA replication are those encoding a DNA polymerase, a DNA ligase (Gp30), and three proteins with helicase activity. The closest homologs to the phage TG1 and ϕR1-RT polymerases are found in Edwardsiella phage PEi20 (accession number BAQ22701.1) and enterobacterial phage RB69 (NP_861746.1), all members of the Myoviridae. Among the helicases, Gp41 (or Dda) and UvsW homologs are involved in the reorganization of stalled DNA replication forks (39). Other putative proteins identified include homologs to the DNA polymerase sliding clamp loader complex Gp44/Gp62, sliding clamp accessory protein Gp45, single-stranded DNA binding protein Gp32, DNA helicase loader Gp59, and Gp61. In phage T4, Gp61 is a primase that interacts with helicase Gp41 to form a helicase-primase complex (or primosome). The primosome, together with the DNA helicase loader Gp59, unwinds the DNA template and primes DNA synthesis on the discontinuous strand. Among the proteins involved in recombination are type II topoisomerases Gp60 and Gp52, the recombination-related endonuclease pair Gp46/Gp47, the RecA-like recombination protein UvsX, and a single-stranded DNA binding protein, UvsY (40). Lastly, among the proteins involved in repair, a DenV homolog and several RNA ligases were identified. DenV is an N-glycosylase UV repair enzyme that excises pyrimidine dimers, the major UV lesions of DNA, while RNA ligases seal breaks in RNA and may also counteract host defense of cleavage of specific tRNA molecules (41).

Nucleotide metabolism.

Class I ribonucleotide reductases are responsible for the interconversion of ribo- to deoxyribonucleotides and are represented by NrdAB or NrdEF, which require oxygen for activity, class II contains NrdJ, and the oxygen-sensitive class III is represented by NrdGH (42). In TG1 and ϕR1-RT, genes coding for the aerobic ribonucleotide reductase complex subunits NrdA, NrdB, and NrdH were identified. Additionally, NrdC genes were also located. Other genes involved in nucleotide metabolism that were identified include those encoding thymidylate synthase (Td), thymidine kinase (Tk), dNMP kinase, dCMP deaminase (Cd), dihydrofolate reductase (Fdr), dCTPase-dUTPase, and the exo-DNase (DexA) and endo-DNase (DenA). A combination of at least some of these genes is required to supplement the intracellular pool of nucleotides for phage DNA and RNA synthesis (41).

Transcription.

Based on the genome maps presented (see Fig. S2 and S3 in the supplemental material), phages TG1 and ϕR1-RT present similar gene arrangements. A search for promoters based on sequence similarity to the host consensus s70 promoter TTGACA(N15-18)TATAAT with a 2-bp mismatch identified 22 probable host promoters in the phage TG1 genome and 24 probable host promoters in the ϕR1-RT genome, which probably function in early transcription (see Tables S5 and S6 in the supplemental material). Additionally, 15 of the putative host promoters are located in the same relative genomic positions within each phage genome. The genomic layout, however, makes it clear that there must be additional promoters functioning to direct the transition from host to viral metabolism. A search for phage-specific promoters using PHIRE (43) and by analysis of sequences of 100 bp in length upstream of each ORF and submitting them to MEME (44) did not yield additional promoters that could be annotated with confidence. A search for putative rho-independent transcription terminators using ARNold (45, 46) yielded 21 putative terminators in the phage TG1 genome (see Table S7 in the supplemental material) and 24 in the phage ϕR1-RT genome (see Table S8). Nevertheless, the presence of phage T4 protein homologs involved in the transcription of late genes, RegA, Gp33, and the sigma factor for late transcription Gp55, suggest that the mechanism for controlling late transcription is similarly complex (41). Likewise, the presence of repressor and translational regulatory protein homologs involved in middle and late transcription, including RegB, DsbA, Alc, MotA, and AsiA, lend further support to this suggestion.

Morphogenesis.

The putative structural proteins of TG1 and ϕR1-RT are homologous to existing phage proteins of the T4 supergroup of viruses (see Table S4 in the supplemental material). Among the putative phage structural genes, the phage head is likely composed of those encoding the major capsid protein Gp23 and the phage capsid vertex protein Gp24. The prohead precursor and scaffolding proteins Gp68 and Gp67, as well as internal head proteins ipIII and ipII, were also identified. Lastly, the head portal vertex protein Gp20 that is connected to the neck and through which DNA enters during packaging and exits during infection was also identified. The whiskers and neck are composed of fibritin (wac) and the head completion proteins Gp13 and Gp14. The tail proteins include the tail sheath terminator Gp3, the tail completion protein Gp15, the tail sheath subunit Gp18, and the tail tube subunit Gp19. Proteins that form the baseplate wedge subunits and tail pins that then go on to associate with the central hub to form the viral baseplate include Gp5, Gp6, Gp7, Gp8, Gp9, Gp10, Gp11, Gp25, Gp27, Gp28, and Gp53. Among these, Gp5 (ORF150) contains a predicted bacteriophage T4-like lysozyme domain (cd00735) or phage lysozyme domain (pfam00959), which aids in penetration through the peptidoglycan layer during the initial infection process. In phage T4, Gp8 and Gp9 connect the long tail fibers (LTFs) of the virus to the baseplate and trigger tail contraction after viral attachment to a host cell, while Gp11 connects the short tail fiber (STF) protein Gp12 (ORF159) to the baseplate (47). The baseplate wedge subunit Gp25 forms a structural component of the outer wedge of the baseplate that has lysozyme activity, evident from the presence of a conserved gene 25-like lysozyme domain (pfam04965). Based on homology and gene synteny, the proteins forming the LTFs in TG1 and ϕR1-RT are composed of the tail fiber proximal subunit Gp34, the tail fiber connector or hinge protein Gp35, the proximal tail fiber protein Gp36, and the distal tail fiber protein Gp37 (47). A variety of chaperones or assembly catalysts involved in morphogenesis were also discovered. Head formation chaperones include the capsid vertex assembly chaperone, the prohead assembly proteins Gp21 and Gp22, and the head assembly chaperone protein Gp31. Chaperones involved in tail formation include the baseplate hub assembly proteins Gp26 and Gp51. Chaperones for tail fiber assembly include gp57A, gp57B, and Gp38.

Host cell recognition elements.

In phage T4, phage tail-associated receptor-binding proteins (RBPs) Gp37 and Gp12 are necessary for host cell recognition, attachment, and initiation of infection. In the phage TG1 and ϕR1-RT genomes, ORF250 codes for putative RBPs of 609 amino acid residues and 503 amino acid residues in length, respectively, sharing 60% overall sequence identity. These proteins also share 40% sequence identity to the distal LTF RBP of Cronobacter phage vB_CsaM_GAP161 (YP_006986537.1) and are homologs to the LTF RBP Gp37 of phage T4 (AJC64544.1). An alignment of these two proteins reveals a high degree of conservation at the N terminus associated with the proximal tail fiber, as well as at the C terminus associated with host recognition (see Fig. S4 in the supplemental material). More specifically, the C-terminal 63 amino acids show 95% sequence identity. Similarly, in the phage TG1 and ϕR1-RT genomes, ORF159 codes for the STF protein Gp12, both of which are 446 amino acid residues in length and almost identical to each other, sharing 95% overall sequence identity (see Fig. S5 in the supplemental material). These proteins are homologous to the STF protein Gp12 of phage T4 (NP_049770.1).

DNA packaging.

In phage TG1, the ORF164 and ORF165/ORF167 genes code for the small (TerS) and large (TerL) DNA packaging subunits, respectively, of a phage terminase protein complex (or holoterminase) that initiates, drives, and terminates translocation of phage DNA into proheads (48). The homologous genes in phage ϕR1-RT are represented by ORF165 (TerS) and ORF166/ORF168 (TerL). Usually, terS and terL are arranged side by side, but in phages TG1 and ϕR1-RT, two ORFs homologous with terL are found. The proteins encoded by ORF165 in TG1 and ORF166 in ϕR1-RT show sequence similarity to the N terminus of the phage T4 TerL. Likewise, the proteins encoded by ORF167 in TG1 and ORF168 in ϕR1-RT show sequence similarity to the C terminus of phage T4 TerL. BLASTX analysis (37, 38) reveals that the terL gene in both phages is interrupted by a transposase (PHA02552). Additionally, during packaging, the DNA ends are also protected against host RecBCD nuclease action by Gp2, the DNA end protector protein (49) identified in phage TG1 and ϕR1-RT as the product of ORF146.

Homing endonucleases.

Homing endonuclease genes (HEGs) are not genuine phage DNA but rather belong to intron-associated selfish DNA elements (50) and are commonly found interspersed throughout Myoviridae genomes (41). Among the HEGs identified in phage TG1, ORF148 and ORF232 exhibit similarity to shortened helix-turn-helix (HN-H) endonuclease genes, and ORF9, ORF43, and ORF66 exhibit similarity to GIY-YIG group I intron endonuclease genes. BLASTX analysis (37, 38) reveals that the gene coding for Gp47 (recombination-related endonuclease II) is divided by ORF66, which contains the HEG. Likewise, the gene coding for UvsX is intersected by ORF43, which contains the HEG. ORF232 also divides the nrdA gene. In phage ϕR1-RT, five HEGs are also found throughout the genome; of these, only ORF148 is homologous to the helix-turn-helix (HN-H) endonuclease gene that is also located in phage TG1 between the genes coding for Gp4 and Gp53. ORF20, ORF51, ORF163, and ORF234 exhibit similarity to the GIY-YIG group I intron endonuclease genes, none of which interrupt or intersect other ϕR1-RT genes.

Lysis.

The final stage of the phage lytic cycle involves the degradation of the bacterial cell wall and release of progeny phages induced by the effect of a pore-producing protein, the holin, and a peptidoglycan-degrading enzyme, the endolysin (51). In TG1, ORF127 and ORF122 in ϕR1-RT each encode an obvious endolysin containing a bacteriophage T4-like lysozyme protein domain (pfam00959) and phage-related muramidase (COG3772). Access of the endolysin to the cell wall occurs through the presence of the holin. Holins are small phage-encoded proteins characterized by the presence of transmembrane domains (TMDs), which accumulate in the cytoplasmic membrane during infection until suddenly, at a specific time, they trigger the formation of lethal lesions, resulting in destruction of the cell wall (51, 52). A search for the TG1 and ϕR1-RT holins revealed that the putative product of their respective ORF253 genes contains a predicted t-holin domain (pfam11031) with 70% identity to the phage holin of enterobacterial phage CC31 (accession number YP_004010117.1). The protein sequences are predicted to contain a single TMD spanning amino acids 30 to 49, as well as a large C-terminal periplasmic domain spanning the amino acid residues from position 50 to the end terminal amino acid at position 218, a characteristic bitopic topology found in the holin proteins of T4-like phages (53) Moreover, as in phage T4, the putative holin gene is separated from the endolysin gene. An additional search for Rz and Rz1 genes coding for transmembrane spanins involved in the disruption of the outer membrane of the host was also conducted based on gene arrangement and membrane localization signals (54). The search revealed two candidate genes, ORF225 and ORF224 in phage TG1 and ORF227 and ORF226 in phage ϕR1-RT, homologous to phage T4 pseT.3 (Rz) and pseT.2 (Rz1), respectively. As in phage T4, the Rz and Rz1 genes are adjacent to each other, arranged with overlapping stop and start codons, and in addition no part of the Rz1 sequence is embedded within the Rz coding region (54). In TG1 and ϕR1-RT, Rz possesses a single amino-terminal TMD, and Rz1 encodes an outer membrane lipoprotein based on the presence of a signal peptidase II (SPII) cleavage site located between amino acid residues 16 and 17, as predicted by LipoP (54, 55). Lastly, the presence of phage T4 homologs to rI-encoded lysis inhibition regulator membrane protein and rIII-encoded lysis inhibitor accessory protein in TG1 and ϕR1-RT suggest the potential for lysis inhibition (LIN) following superinfection (56, 57).

Phylogeny of TG1.

It is interesting to note that very similar bacteriophages with an overall DNA sequence identity of 91% were isolated from very different locations and sources, as phages TG1 and ϕR1-RT were isolated in Canada from pig manure and in Finland from raw sewage, respectively. Moreover, less than 34% overall DNA similarity exists with their closest neighbors within the Myoviridae (see Table S9 in the supplemental material). The relatedness of these two phages was further explored using progressiveMauve (Fig. 2) (58, 59), CoreGenes (60, 61), which the Bacterial and Archaeal Virus Subcommittee of the International Committee on Taxonomy of Viruses (ICTV) has used extensively to compare the proteomes of viruses, and phylogenetic analysis of their whole-genome sequences (see Fig. S6) and their large terminase subunit protein sequences (Fig. 3). It is evident from phylogenetic analyses that TG1 and ϕR1-RT form a distinct taxonomic clade among their closest neighbors. Based on these observations and using a 95% DNA sequence identity as the criterion for demarcation for a species, a new genus named Tg1virus, with phages TG1 and ϕR1-RT (R1RT in the ICTV database) as member species, was proposed to the ICTV (approved and ratified in 2016).

FIG 2.

ProgressiveMauve alignment of phage TG1 and ϕR1-RT. The genome of ϕR1-RT (accession number HE956709) is shown at the top, and that of TG1 (accession number KP202158) is shown at the bottom of the figure. The degree of sequence similarity between regions is given by a similarity plot within the colored blocks, with the height of the plot proportional to the average nucleotide identity. Below these are illustrated the phage genes as outlined boxes on the plus (above horizontal) and minus (below horizontal) strands.

FIG 3.

Phylogenetic analysis of the large terminase subunit protein sequences of phages TG1, ϕR1-RT, and related bacteriophages. The phylogenetic analysis was constructed using “one click” at phylogeny.fr (at http://phylogeny.lirmm.fr/phylo_cgi/index.cgi), with MUSCLE used for multiple alignment and PhyML used for phylogeny (85).

Growth curves.

In order to study the efficiency of phage infection at different temperatures, bacterial growth after phage infection with ϕR1-RT was measured. The host bacterial strain was grown at selected temperatures with the addition of different phage stock dilutions. Bacterial growth was followed for 3 days at 4°C, 2 days at 10°C and 16°C, and 1 day at 22°C and 37°C. Lysis of the bacterial cultures was observed at 4°C, 10°C, 16°C, and 22°C, whereas at 37°C, the bacteria were not significantly affected even with the highest initial phage concentrations (Fig. 4A). The onset time of lysis depended on the temperature and initial phage titer. At 4°C, the bacterial culture started to lyse after 56 to 60 h. At 10°C, the culture infected with the highest phage titer started to lyse after 16 h, and the culture infected with the lowest phage titer started to lyse after 24 to 28 h. The corresponding times for 16°C and 22°C were 6 and 12 h. While the lysis at 10°C and 16°C was complete, at 22°C, strong regrowth took place after the initial lysis. At 4°C, the 3-day incubation time was not long enough to follow the lysis to completion. Under all tested conditions, negative (medium only) controls showed no increase in absorbance, whereas the positive (bacteria only) controls presented the normal bacterial growth pattern.

FIG 4.

Phage ϕR1-RT does not propagate at 37°C. (A) Growth curves of Y. enterocolitica infected with ϕR1-RT. Bacteria were cultured with different concentrations of phage particles in LB at 4°C, 10°C, 16°C, 22°C, and 37°C. Each graph represents the average results for five replicates. Note the different scales used for the x axis at different temperatures. (B) Analysis of ompF gene expression (left) and protein abundance (right) at different temperatures. The mean expression levels of the ompF gene were obtained from RNA sequencing analysis. The production levels of the OmpF protein were obtained from normalized mean spectral values for the proteins detected by LC-MS/MS analysis. Error bars represent the calculated standard deviations.

Transduction.

To study the transducing potential of ϕR1-RT, transduction of the Km^r and urease-negative phenotype of strain YeO3-hfq::Km to the Km^s and urease-positive wild-type strain 64741/76 was assayed. Repression of urease activity is one of the phenotypes of the hfq mutant (62) and may be used to confirm the transduction of the hfq::Km allele. The transduction assays were performed in 10 parallel tubes using an MOI of 0.006. A total of 6.6 × 10⁸ PFU from the transducing lysate resulted in a total of 3 Km^r urease-negative colonies that were confirmed by PCR. From this, the calculated transduction frequency in the experiment was 4.5 × 10⁻⁷ transductants per PFU.

Identification of the phage receptors. (i) Pseudolysogeny.

As the LPS and protein profiles of the phage-resistant mutants YeO3-ϕR1-RT-2, YeO3-ϕR1-RT-7, and YeO3-ϕR1-RT-9 did not differ from those of the wild-type bacteria (data not shown), the genomic DNA of the mutant and wild-type strains were sequenced. The de novo assembly results showed that the total scaffold sizes of the assembled genomes of the three mutants were ∼165 to 173 kb larger than that of the wild-type parental strain (see Table S10 in the supplemental material). This suggested that the mutants carry extra DNA very similar in size to the phage ϕR1-RT genome (168,809 bp). This immediately raised the possibility that the phage had lysogenized the host and would reside as a prophage. For all three draft genomes, the phage genome sequence was indeed identified, and in all it formed the fourth largest scaffold (4.1) with almost identical sizes (see Table S10). Significantly, in all three cases, the scaffold sequences were 100% identical to the phage ϕR1-RT sequence without any flanking host sequences, suggesting that the phage genome resided in these bacteria as an autonomous replicating unit in a state known as pseudolysogeny. Such a state has been described for T4-like phages (63).

(ii) Transposon insertion library screening.

As selection of spontaneous phage-resistant mutants seemed to favor pseudolysogeny, we decided to use a different approach. A Cat-Mu transposon library of strain YeO3-R1 (26) was exposed to the ϕR1-RT for 2 h, and the surviving phage-resistant mutants were grown on LA-Clm plates. The recovered colonies were tested to be true ϕR1-RT-resistant mutants. In order to exclude pseudolysogens, the clones were screened with an ϕR1-RT-specific PCR, and negative ones were further analyzed by a Cat-Mu-specific arbitrary PCR to identify the Cat-Mu insertion site (26). For four of the candidates, the transposon insertion site was identified as gene Y11_04441 of the Y. enterocolitica O:3 strain Y11 genome (NC_017564.1). In the strain Y11 genome, the gene was annotated to encode the outer membrane porin OmpC, but in all other Y. enterocolitica genomic sequences, the gene was annotated to encode OmpF, so we opted to use OmpF. To confirm that OmpF is the ϕR1-RT receptor, one of the YeO3-R1-Cat17 mutants was complemented with the wild-type ompF gene either in trans with plasmid pTM100_OmpF or in cis by suicide plasmid pSW25T_OmpF. Both of these approaches resulted in regaining the phage sensitivity, thus confirming that OmpF serves as the ϕR1-RT receptor.

The LPS inner core heptose region functions as a receptor.

Adsorption experiments were carried out to study the ability of ϕR1-RT and TG1 to interact with Y. enterocolitica O:3 derivatives differing mainly in their LPS compositions (Fig. 5). A short 5-min adsorption time was used, as it produced the highest resolution between the strains. A general observation was that TG1 adsorbed faster than ϕR1-RT. Both phages showed negligible adsorption to YeO3-c-R1-Cat17, the ompF mutant strain, and adsorbed well to both ompF-complemented strains. Both phages showed reduced but clear adsorption to the pseudolysogen, indicating changes in abundance or exposure of the phage receptor(s). Finally, the adsorption to the inner core mutants decreased with the truncation of the core oligosaccharide, suggesting that the inner core heptoses are part of the secondary receptor (Fig. 5).

FIG 5.

Phages ϕR1-RT and TG1 use OmpF and the LPS inner core heptose region of Y. enterocolitica O:3 as receptors. (A) Adsorption experiments were performed with different LPS and ompF mutants, with the complemented strains, and with the pseudolysogen. All strains are OmpF positive with the exception of YeO3-c-R1-Cat17. The TG1 and ϕR1-RT adsorptions to the bacteria at 5 min are shown as percentages of residual PFU. Error bars indicate standard deviations. The no-bacteria control (LB) and strains used for adsorptions are indicated below the columns. The LPS chemotypes of the strains are indicated at the top of the columns. (B) Schematic structures of the Y. enterocolitica O:3 LPS molecules of different chemotypes (86). Note that Y. enterocolitica O:3 carries simultaneously the S and Ra types of LPS molecules. This is indicated in panel A by a plus sign. O-ag, O-antigen or O polysaccharide; OC, outer core hexasaccharide; IC, inner core; LA, lipid A.

Temperature dependence of ompF expression.

We then wondered whether the temperature-dependent sensitivity of Y. enterocolitica O:3 could be due to ompF regulation. The expression of ompF under different growth temperatures was analyzed from RNA sequencing and quantitative proteomics (liquid chromatography-tandem mass spectrometry [LC-MS/MS]) data. The transcriptomic data showed an inverse correlation between the expression of ompF and the temperature of incubation (Fig. 4B). Consistently, the quantitative proteomics demonstrated a much higher abundance of the OmpF protein in the 22°C sample than in the 37°C sample, where the abundance barely exceeded the threshold of identification (Fig. 4B).

In vitro expression of the LTF protein Gp37 of phage TG1.

In this study, coexpression with the phage-encoded chaperones Gp38 (required for oligomerization) and Gp57A, which is also thought to participate in assembly (64, 65), was utilized in an attempt to synthesize the native form of distal LTF protein of phage TG1, as previously described for the production of Gp37 from phage T4 (28). SDS-PAGE demonstrated that an oligomer of approximately 210 kDa was obtained when Gp37 was coexpressed with Gp38 in a bicistronic plasmid (pCDF Duet-1 Gp37-Gp38) or when this same plasmid was coexpressed with Gp57A (see Fig. S7, lanes 3 and 5, in the supplemental material). Under reduced conditions, Gp37 appears as a monomer of approximately 70 kDa in size (see Fig. S7, lanes 4 and 6). This estimate is close to the predicted molecular mass of the recombinant phage TG1 Gp37 determined via MS to be approximately 68.050 kDa. Peptide mass fingerprinting confirmed the identity of the protein (see Fig. S8). Based on the protein expression results obtained, it appears that in phage TG1, only the Gp38 chaperone is essential and the general chaperone Gp57A is not required for in vitro protein folding of Gp37 as has been reported for phage T4 (28). The formation of higher-molecular-weight oligomers of phage TG1 Gp37 is consistent with previous reports that describe RBPs of phages present as homotrimers in solution migrating in the SDS-PAGE with a mobility that corresponds to that of oligomeric forms (28, 66–68).

Confirmation of host binding specificity.

Host binding specificity was then tested through the immunolabeling of bacterial cells with phage TG1 LTF protein Gp37, followed by detection with anti-His₆ antibodies and DyLight 488 conjugated secondary antibodies. Consistent with the temperature-dependent infection of phage TG1, the application of the LTF protein Gp37 to Y. enterocolitica cells showed decoration of the surfaces of Y. enterocolitica O:3, O:5,27, and O:9 cells when these were grown at 25°C but not at 37°C (Fig. 6). Notably, binding was more apparent near the apex of the cells, which is also reported to occur in other phages, such as λ, T4, T7, KVP40, and ϕA1122, preferentially infecting cells at the poles (69).

FIG 6.

Confocal immunofluorescence microscopy images of Y. enterocolitica cells after incubation with LTF protein Gp37 derived from phage TG1. Gp37 decorates the cell surfaces of Y. enterocolitica strain K14 of serotype O:9 (a), Y. enterocolitica strain gc815-73 of serotype O:5,27 (c), and Y. enterocolitica strain 6471/76 of serotype O:3 (e) grown at 25°C, whereas Y. enterocolitica strain 8081 of serotype O:8 (g) does not show cell decoration with Gp37. Images similar to that shown in panel g were observed when the same strains were grown at 37°C. Differential interference contrast microscopy images of panels a, c, e, and g, are shown in panels b, d, f, and h, respectively. The scale bars represent sizes in micrometers.

DISCUSSION

Among bacteriophages, the C termini of RBPs involved in ligand interactions usually exhibit considerable sequence divergence, thus providing diversity in host specificity. In the case of ϕR1-RT and TG1, the high sequence identity at the C termini of their LTF and STF proteins may account for the striking similarity in virulence of these two phages for Y. enterocolitica. Notably, phage ϕR1-RT shows virulence to strains of the same serotypes as phage TG1. Based on adsorption experiments, the outer membrane protein OmpF and the inner core heptosyl residues of the LPS serve as phage receptors for phage TG1 and ϕR1-RT. It is worth noting, however, that the E. coli strain DH10B/pTM100_OmpF was not sensitive to ϕR1-RT. We reasoned that this could be due to poor expression of Y. enterocolitica OmpF in E. coli or, more likely, that the LPS inner core, known to be used by T4-like phages as the secondary receptor (70, 71), was not compatible. The inner core structures of E. coli and Y. enterocolitica differ substantially, potentially explaining this result.

Multiple lines of evidence suggest that OmpF is the primary host range determinant for these two bacteriophages. First, a multiple alignment of OmpF amino acid sequences of Y. enterocolitica (from a BLASTP search of sequence databases using the O:3 OmpF sequence as the query) suggests that the restricted host range of these phages among Y. enterocolitica serotypes could be due to OmpF. The alignment provided a distribution of conserved amino acid residues and the presence of regions with high and low homologies, which coincide with eight transmembrane domains and eight “external” loops, respectively, of the topology of the OmpF porin from E. coli (72, 73). The search and alignment of the sequences (see Fig. S9 in the supplemental material) revealed that the OmpF sequences of the ϕR1-RT-sensitive serotypes are 100% identical. The most dramatic differences between the serotypes map to loop 4. In the alignment, the closest match to the O:3 sequence is the serotype O:7,8,19 OmpF sequence, which is 96% identical to O:3 and may still be sensitive to ϕR1-RT; in it, the loop 4 sequence differs least, while in others, differences are bigger and also accumulate in other loops, mainly in loops 5, 6, and 7 (see Fig. S9). The porin loops are plausible binding sites for bacteriophages, as demonstrated by the interaction of E. coli OmpF and K20 phages, which bind to the L5, L6, and L7 external loops (74–76). Thus, it is likely that the loop 4 sequence is targeted by the ϕR1-RT or TG1 receptor binding proteins, but experimental evidence is necessary to confirm this. Second, RNA sequencing and quantitative proteomics data, the analysis of growth curves of Y. enterocolitica infected with ϕR1-RT at various temperatures (4°C to 37°C), and results of phage host range analysis conducted at 25°C and 37°C clearly indicate that the failure of ϕR1-RT and TG1 to infect Y. enterocolitica O:3 at 37°C is due to the strong repression of the ompF gene. The results showing the temperature-dependent expression of OmpF also agree with a previous study, where two-dimensional gel electrophoresis of whole-cell proteins of Y. enterocolitica cultured at 25°C and 37°C suggested that OmpF is downregulated when the bacteria were cultured at 37°C (77). Consistent with this observation, the application of immunolabeled phage TG1 receptor binding protein Gp37 to Y. enterocolitica cells showed decoration of the surfaces of Y. enterocolitica O:3, O:5,27, and O:9 cells when these were cultured at 25°C but not at 37°C. The decoration of the cell surface agrees with a high-level expression of this major outer membrane protein class, depending on the bacterial species and the environmental conditions, which can reach about 10⁴to 10⁶ copies per cell (74). It is reasonable to suggest then that the phage TG1 distal LTF protein Gp37 (and by extension, its homolog in ϕR1-RT) is specifically involved in binding to OmpF while presumably the STF protein Gp12 binds to the inner core of LPS, as is reported to occur in other T even phages as a secondary receptor (70, 71).

The in vitro temperature-dependent infection of these two highly related phages questions their potential use as biocontrol or therapeutic agents, as has been suggested for the temperate Yersinia phage PY100 (18, 78). On the other hand, it is not known whether the ompF gene is expressed in vivo, justifying further studies toward finding that out. However, due to their marked specificity for the epidemiological relevant Y. enterocolitica serotypes O:3, O:5,27, and O:9, these phages may prove useful for diagnostic purposes. In addition, the successful synthesis of the LTF of phage TG1 opens up the possibility of its use as a probe, as well as for the production of suitable amounts of protein for X-ray crystallography to elucidate its atomic structure or cocrystallization with its receptor OmpF to shed light on specific host cell receptor-virus interactions.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.