A Tailspike with Exopolysaccharide Depolymerase Activity from a New Providencia stuartii Phage Makes Multidrug-Resistant Bacteria Susceptible to Serum-Mediated Killing

Phage vB_PstP_Stuart (further referred to as Stuart) specifically infects four (PS#3, PS#5, PS#11, and PS#15) out of 15 P. stuartii strains with five different phenotypic patterns of antibiotic resistance (Table 1). The infected strains represented three of these patterns (numbered as 3, 4, and 5 in Table 1). However, strains representing patterns 3 and 5 included phage Stuart-sensitive as well as resistant isolates, indicating further phenotypic differentiation of these strains. No lysis of P. rettgeri and P. alcalifaciens or Morganella, Proteus, and Citrobacter species representatives closely related to P. stuartii was observed (Table 1). Phage Stuart appeared to form clear and large plaques (2 to 3 mm in diameter) on lawns of sensitive bacteria. On the lawns of P. stuartii PS#3, PS#11, and PS#15 cells, but not on the lawn of PS#5, the plaques were surrounded by hazy rings that expanded over time. This indicates that phage Stuart can encode a depolymerase, which specifically degrades exopolysaccharides (EPSs) of certain P. stuartii strains (Fig. 1).

Transmission electron micrograph (TEM) images illustrate that phage Stuart belongs to the Podoviridae family of the order Caudovirales (Fig. 2A). Its virion (n = 7) consists of an icosahedral head (63 ± 2 vertex to vertex, 58 ± 2 width) and a noncontractile tail with short tail fibers (head edge to tail end 12 ± 3 nm in length) typical of Podoviridae. The latent period of phage Stuart is 25 min (5 min of the initial adsorption plus the 20-min period without an increase in phage titer), and a burst size is 101 phage particles per infected cell, as indicated by the analysis of the one-step growth curve (see Fig. S1 in the supplemental material).

Phage Stuart has a 41,218-bp genome (Fig. 2B) with direct terminal repeats (TRs) of 345 bp and a GC content of 42.9%. The ends of TRs vary by a few base pairs between DNA molecules of particular virions as indicated by the comparison of single Illumina reads of end fragments. The structure of TRs is characterized by the GC-rich stem-loop region that is followed by the G-rich region and the tandem repeats of the TAAGTCCAG sequence followed by one imperfect and incomplete repeat (TAAGCCAC). The number of repeats in the population of virion DNA molecules varies from 3 to 5 as determined based on the alignment of single Illumina reads of the repeat regions.

Searches for related phages with BLASTn revealed that phage Stuart has overall nucleotide identity (query coverage × percent identity) of about 48% to 49% with four Proteus phages of the Autographivirinae subfamily (vB_PmiP_RS8pmA, PM16, vB_PmiP_RS1pmA, and PM 75; GenBank accession numbers MG575419, NC_027342, MG575418, and NC_027363, respectively). Such low levels of similarity between phage Stuart and its closest relatives (far below the genus demarcation criterion of 95% [23]) allow one to classify phage Stuart to a new genus within the Autographivirinae subfamily. DNA identity of phage Stuart to other members of the Autographivirinae subfamily did not exceed 11%. PROmer alignment demonstrated a decreased order of homology of phage Stuart with Proteus phages PM16 and PM 75, followed by prototypic Klebsiella phage KP34 (Drulisvirus genus) and Pseudomonas phage phiKMV (Phikmvvirus genus) (see Fig. S2 in the supplemental material). Phage Stuart shares 34, 19, and 13 genes with PM16 or PM 75, KP34, and phiKMV, respectively.

Although phage Stuart is predicted to encode 52 proteins, only 26 of them can be assigned function related to DNA replication, repair and modification (e.g., DNA primase, DNA polymerase), cell lysis (endolysin, holin, and spanin), and DNA packaging and structural protein (e.g., major capsid, portal protein) (see Table S1 in the supplemental material). No genes that could encode tRNAs were found. Of particular interest is the atypical cell lysis cassette region, which can encode u-spanin (gp49), pin-holin (gp50), and signal-arrest-release (SAR) endolysin (gp51) proteins (Fig. 2B; see also Table S1). This topology is similar to that of Klebsiella phage KP34 and Proteus phages PM 75 and PM16, although the endolysins of the last two phages lack SAR sequence peptides. However, the organization of the phage Stuart holin-independent lytic system differs from that of phiKMV-like phages (KMV44, KMV45, and KMV46), which encode the pin-holin, SAR endolysin, and i-/o-spanin (24). The Stuart virion contains at least twelve proteins (gp32, gp34, gp36, gp37, gp39 to gp42, gp44, gp45, gp48, and gp52), as identified by using tandem mass spectrometry analysis (see Fig. S3 in the supplemental material). Functions of the majority of them can be predicted. Only gp48 has no homologs.

gp52 is a 604-amino-acid virion protein that was identified by mass spectrometry with 31 unique peptides and a sequence coverage of 69% (see Fig. S3). The predicted molecular mass is 66.86 kDa. Sequence comparisons of gp52 with BLASTp and HHpred show a weak similarity to the putative tailspikes of Acinetobacter phage IME-AB2 and Abp1 (32% amino acid identity) and to proteins of unknown function of Acinetobacter spp. The amino acid sequence of gp52 has the following two conserved Pfam motifs: the head-binding motif (PF09008.9; E-value, 4.5e-12) and the pectate lyase 3 motif (PF12708.6; E-value, 3.7e-8). Additionally, the predicted structure of gp52 is similar to the structure of the Acinetobacter phage vB_AbaP_AS12 tailspike (Protein Data Bank accession number 6EU4) (Fig. 3A).

As phage tail-associated depolymerases harboring the pectate lyase domains are found to degrade bacterial capsules of Acinetobacter and Klebsiella hosts (14, 25), the phage Stuart tailspike was expressed in a His-tagged form in E. coli, purified, and tested for the ability to form halo zones on the layers of Providencia species cells. In a spot-on-lawn test, the tailspike produces opaque halos from 0.007 to 7 μg (Fig. 3B) and is active against three isolates (PS#3, PS#11, and PS#14). Strain PS#5 is therefore susceptible to phage Stuart infection but appeared to be insensitive to the tailspike activity. In extracted exopolysaccharide (EPS) degradation assays, the tailspike produced reducing ends, indicating that it has EPS-degrading activity (Fig. 3C). Given the unknown number of polymers present in the EPS, we correlated the tailspike EPS-degrading activity with 0.5 μM glucose equivalent reducing ends. We further demonstrate that the tailspike can degrade EPS in the presence of 5 mM EDTA, meaning that the enzyme appears to act independent of divalent cations.

To gain insights into the tailspike structural features, we analyzed its secondary structure. Circular dichroism (CD) spectroscopy revealed that the protein adopts a well-folded conformation with rich β-sheet structures with a negative dichroic minimum at 215-nm and a positive dichroic maximum at 195-nm characteristic peaks (Fig. 4A). CD spectrum deconvolution showed that the tailspike folds in 38% of β-strands, 7% of α-helices, 22% of turns, and 32.9% of unordered structures, similar to other virion receptor-binding proteins (26, 27). The estimate of β-structures (43%) predicted by PSIPRED is in agreement with a predominance of β-strands in the tailspike (28). To monitor protein secondary structure transitions induced by thermal stress, melting curves were recorded following the CD signal at 215 nm as a function of temperature (Fig. 4B). The tailspike demonstrated a sigmoidal thermal denaturation profile, reflecting protein unfolding with a melting temperature (T_m) of 72°C. Finally, the oligomerization state was also investigated running polyacrylamide gels (Fig. 4C). During electrophoresis in a polyacrylamide gel under nondenaturing conditions, the native tailspike migrated as a protein of molecular mass of approximately 201 kDa, which was three times higher than the molecular mass of tailspike estimated based on its migration in the denaturing gel (Fig. 4C) and based on the mass spectrometry analysis (67 kDa) (see Fig. S3). This indicates that the tailspike forms a putative trimer.

As the tailspike was found to degrade the bacterial EPS, we further explored its role as a receptor-binding protein. For this, we performed phage adsorption assays with phage Stuart-insensitive (PS#1 and PS#2) and -sensitive (PS#3, PS#5, PS#11, and PS#15) strains pretreated with phosphate-buffered saline (PBS) or the tailspike (Fig. 5). For PS#1 and PS#2, no adsorption to both PBS- and tailspike-treated cells was detected. For PS#3, PS#11, and PS#15, results showed that 90% versus 7% (for PS#3), 78% versus 16% (for PS#11), and 79% versus 8% (for PS#15) of phage particles were adsorbed for PBS- and tailspike-treated cells, respectively. In the case of the PS#5 strain, which is sensitive to phage Stuart but not to the tailspike (Table 1), 91% versus 90% of phage particles were adsorbed to the PBS- and tailspike-treated cells, respectively. Conceivably, phage Stuart uses a different receptor on PS#5 cells than that on PS#3, PS#11, and PS#15 cells, and this receptor is not susceptible to tailspike-mediated degradation.

To test the antivirulence properties of the tailspike, we used a human serum model. The serum killing effect was evaluated by mixing serum of different healthy subjects with insensitive (PS#1 and PS#2) and sensitive (PS#3, PS#5, PS#11, and PS#15) strains to phage Stuart and treating them with PBS or the tailspike (Fig. 6A). The number of colony-forming PS#1 and PS#2 cells was reduced by >4 logs (below the detection limit of 10 CFU/ml) or slightly reduced (0.5 logs), and the addition of the tailspike had no effect. PS#3, PS#11, and PS#15 strains could resist serum. However, when coadministered with the tailspike, the numbers of their cells were reduced by >4 logs (below the detection limit of 10 CFU/ml). Strain PS#5 was killed by the serum and was, therefore, not tested with the tailspike. In all cases, heat-inactivated serum and the tailspike had no antibacterial effect, as no decrease of CFU per milliliter was observed. Our data clearly demonstrate that the tailspike can significantly complement the serum killing in vitro when mixed with cells of sensitive strains. The serum-complementing activity of tailspike depends on its ability to degrade the EPS of sensitive strains.

While treatment with phage can cause the overgrowth of a population with phage-resistant cells, the treatment with the tailspike, which is neither a bactericidal nor bacteriostatic agent, should not impair bacterial growth and should not select for resistance. However, rare random mutants of changed EPS, resistant to degradation by the tailspike, could be present in a bacterial population. As expected, the impact of phage Stuart and the tailspike on the growth of sensitive bacteria and on the emergence of phage-resistant cells varied significantly (data not shown). In the absence of phage Stuart and the tailspike, the optical density of cell culture, grown with shaking at 37°C, increased from 0.1 to 1.0 in 4 h. In the presence of tailspike, the growth rate of culture was similar, while in the presence of phage Stuart, the culture reached an optical density of 1.0 only after 8 to 9 h. Colonies of 20 random cells isolated from cultures upon 72 h incubated with phage Stuart produced cultures insensitive to phage Stuart and to the tailspike. In contrast to that, all 20 colonies obtained from cells of the tailspike-treated cultures remained sensitive to both phage Stuart and the tailspike. Clearly, the tailspike is superior to the phage in that it cannot select resistant phenotypes.

To further support these observations, cells of ten PS#3 colonies obtained after challenge with the tailspike were treated with serum and PBS or with serum and tailspike (Fig. 6B). Cells of all of these colonies remained sensitive to the tailspike in the drop test and were not vulnerable to killing by serum in the absence of tailspike. However, the presence of tailspike in the mixture of cells with serum resulted in the reduction of the number of viable cells by >4 logs (below the detection limit of 10 CFU/ml). Taken together, the tailspike is able to repeatedly sensitize P. stuartii cells to serum even if they are taken from cultures that have been previously challenged by the tailspike.

Clinical specimens of Providencia species strains, including 15 strains of P. stuartii (PS#1 to PS#15), 10 of P. rettgeri (PS#1 to PS#10), and 1 of P. alcalifaciens (PA#1) (Table 1), were obtained from the collection strains of Hospital of Braga (Portugal). They were collected from different human subjects (male and female) and isolation sources (urine, bronchial aspirate, expectoration, cutaneous exudate, and pus) at various times during a period of 3 months. Strains were isolated using CLED (cysteine-, lactose-, and electrolyte-deficient) selective medium. Microbial identification and antibiotic susceptibility testing were performed with Vitek 2 (bioMérieux) or WalkAway (Beckman Coulter) using MIC breakpoints according to CLSI guidelines (M100-S24) (42). Of these, phage Stuart insensitive (PS#1 and PS#2) and sensitive (PS#3, PS#5, PS#11, and PS#15) strains were frequently used in this study. Strain PS#3 was used to propagate the phage Stuart. Additionally, the following collection strains from Salmonella Genetic Stock or American Type Culture Collection centers were used: Citrobacter freundii EC592, Morganella morganii (CDC 4195-69), Proteus mirabilis CECT 4101 (ATCC 14153), Proteus vulgaris CECT 174 (ATCC 6380), and Escherichia coli CECT 432 (ATCC 21026). All strains were grown aerobically at 37°C in Trypticase soy agar (TSA) (Oxoid) or Trypticase soy broth (TSB) (Oxoid) medium and used to study the phage Stuart and the recombinant tailspike activity spectra. E. coli BL21(DE3) was used as a plasmid host to produce the recombinant tailspike. This strain was grown at 37°C in Luria broth (LB) (NZYTech) medium supplemented with kanamycin (50 μg/ml).

Phage Stuart was isolated and purified exactly as described elsewhere (43). In brief, the enriched sewage samples were tested for the presence of phages by their plaque forming capacity on P. stuartii strain lawns. The phage was screened for activity using the spot-on-lawn test and soft agar overlay method. Phage plaques were streak-purified three times before producing high-titer suspensions by soft agar overlay and polyethylene glycol (PEG) purification methods.

The phage Stuart one-step growth curve experiment to determine the latent period and burst size was performed exactly as previously described (25). Briefly, mid-exponential phage Stuart host cells (PS#3) were harvested by centrifugation (7,000 × g, 5 min, 4°C), suspended in TSB, and mixed with a phage suspension (0.001 PFU per cell). Mixed samples (10 ml) were incubated at 37°C for 5 min and centrifuged (7,000 × g, 5 min, 4°C), and the pellet was suspended in 10 ml of TSB and further incubated at 37°C under agitation. After 5, 10, 15, 20, 25, 30, 40, 50, and 60 min, the phage was titrated (PFU/ml). Data were normalized by dividing phage concentration at each point by its initial concentration. Phage burst size was calculated based on the phage titer of the first plateau. Phage latent period was determined as the time until the beginning of phage titer increase.

Transmission electron micrographs (TEM) of phage Stuart were captured using a JEM-1400 microscope (JEOL, Tokyo, Japan) exactly as previously described (44). Lysates were centrifuged (1 h, 25,000 × g, 4°C) to pellet phages and then washed twice with tap water. Subsequently, the suspension was deposited on carbon-coated copper nickel grids for 2 min and stained with 4% uranyl acetate (Agar Scientific) for 30 s. The virion dimension data present mean ± standard deviation from 7 observed phage particles from different TEM.

Phage Stuart genomic DNA was purified using phenol-chloroform-isoamyl alcohol extraction as described elsewhere (45). The library of DNA fragments for sequencing was prepared by mechanical shearing with the use of a NEBNext Ultra DNA library prep kit and sequenced using the Illumina MiSeq platform (2 × 150 bp) at VIB Nucleomics Core (Belgium). Quality-controlled trimmed reads were de novo assembled into a single contig using CLC Bio Genomics Workbench v7.0 (Aarhus, Denmark) with average coverage of 5,806×. An additional library of large phage Stuart DNA fragments was prepared and sequenced using the Oxford nanopore MinION technology (435 fragments up to 12 kb; average, 1,713 bp) to clarify the ambiguous assembly pattern of sequence reads at the regions of putative virion DNA termini. Predicted virion DNA end fragments were separated by digestion with NcoI and PvuI isolated from an agarose gel (with the use of PureLink quick gel extraction kit; Thermo Fisher Scientific) and used as templates for sequencing with the following primers: 5′-CAACCTACTGTGGATAACTCTGT and 5′-AAGTCGGCCCACCTGTAAGC (left end), and 5′-GGTATAGGCCTTGTGTTAGTTTAG and 5′-AGAGTGGGGTAAGCCGTTCG (right end).

Open reading frames (ORFs) were predicted using Glimmer (46) and manually inspected. A search for putative tRNA-encoding genes was performed with the use of tRNAscan-SE (47). Identification of homologs and functions of proteins encoded by the ORFs were made by BLAST algorithms available at the National Center for Biotechnology Information (NCBI). Protein sequences were screened with Phobius (48), TMHMM (49), and SignalP (50) to predict transmembrane domains and signal peptides. Lipo1.0 (44) was used to identify lipoprotein signal peptides (51). Potential promoters and intrinsic (Rho-independent) terminators were predicted with the use of phiSITE (52), PhagePromoter (53), and ARNold (54). Whole-genome comparisons were made with BLAST algorithms against the nonredundant database using progressiveMauve (55), CoreGenes (56), and PROmer/MUMmer (available at http://mummer.sourceforge.net). Results of comparisons were visualized with Easyfig (57) and Circos (58).

For the analysis of virion proteins, phage Stuart was purified with the PEG method. Phage was added to a culture of PS#3 cells at mid-exponential phase and incubated for 8 h. Subsequently, NaCl at 1 μM was added to the suspension, the solution was incubated for 1 h in ice, and was then centrifuged (10 min, 11,000 × g, 4°C). PEG 6000 (10% [wt/vol]) was added to phage lysates, and the mixture was incubated for 16 h at 4°C. After centrifugation (10 min, 9,000 × g, 4°C), the pellet was suspended in SM buffer (100 mM NaCl, 10 mM MgSO₄, 50 mM Tris-HCl, pH 7.5) and chloroform (1:4 [vol/vol]) was added. After centrifugation (15 min, 3,000 × g, 4°C), the upper phase was collected and filtered through a 0.22-μm Whatman polyethersulfone (PES) membrane, and the phage was titrated.

Using a high titer of purified phage suspensions (10¹⁰ PFU/ml), structural proteins were precipitated by chloroform-water-methanol extraction (1:1:0.75, vol/vol/vol) and suspended in loading buffer (1% SDS, 6% sucrose, 100 mM dithiothreitol, 10 mM Tris, pH 6.8, 0.0625% [wt/vol] bromophenol blue). After heating at 95°C, proteins were separated in a 12% polyacrylamide gel by electrophoresis (SDS-PAGE) and stained using GelCode Blue Safe protein stain (Thermo Scientific, Waltham, MA, USA). The entire gel lane containing separated proteins was cut into 24 distinct slices, trypsinized according to the protocol of Shevchenko et al. (59), and analyzed by nanoliquid chromatography-electrospray ionization-tandem mass spectrometry (NanoLC-ESI-MS/MS). Peptides were identified with SEQUEST v1.4 (Thermo Fischer Scientific) and Mascot v2.5 (Matrix Sciences) and a database containing all possible phage Stuart ORFs as described previously to validate gene predictions (60).

The tailspike encoding gene of phage Stuart (gp52) was amplified from the genomic DNA with KAPA HiFi polymerase (Kapa Biosystems) and primers 5′-CATATGGCAAAAGTAAATAGAACTGTAC-3′ and 5′-AAGCTTTTAATTAATCTCGACTTTACGCC-3′, which contained NdeI and HindIII restriction sites, respectively (underlined). The amplicon was purified (DNA Clean & Concentrator-5k; Zymo Research) and digested with NdeI and HindIII. Plasmid pET28a, which contains a sequence encoding a His tag upstream of its cloning site, was digested with NdeI and HindIII, purified (DNA Clean & Concentrator-5k; Zymo Research), and used to ligate the digested pET28a with the help of a T4 ligase (NEB). DNA of ligation mixture was electroporated into E. coli BL21(DE3)-competent cells. The presence of the correct insert in the recombinant plasmid (pET28a-tailspike) was confirmed by Sanger sequencing (GATC Biotech). E. coli BL21(DE3) cells harboring pET28a-tailspike were grown in LB medium supplemented with kanamycin (50 μg/ml) until the exponential phase of growth (optical density of 0.5 at 620 nm), and isopropyl-β-d-thiogalactopyranoside (IPTG) was added to a final concentration of 0.5 mM to induce the expression of the cloned gene. After induction for 18 h at 16°C, cells were harvested by centrifugation (9,500 × g, 30 min), suspended in lysis buffer (50 mM Tris, pH 8.0, 250 mM NaCl, 5 mM MgCl₂), and disrupted by 3 freeze-thawing cycles (−80°C/room temperature). Keeping the sample in ice, cells were further disrupted by sonication for a total period of 5 min (with 30-s pulses interspersed with 30 s resting on ice). Clear supernatants collected after centrifugation (9,500 × g, 30 min, 4°C) were subjected to immobilized metal ion affinity chromatography (HisPur Ni-NTA resin; Thermo Fisher Scientific) using the stepwise increasing imidazole concentrations in the washing and elution buffers (20 mM NaH₂PO₄, 0.5 M NaCl/NaOH, pH 7.4, 25 to 250 mM imidazole). Eluted protein was dialyzed against 20 mM PBS, pH 7.0, and quantified with the Pierce BCA protein assay kit (Thermo Fisher).

The tailspike was characterized using the following two distinct methods: (i) spot test and (ii) measuring the reducing end release during the degradation of extracted exopolysaccharides (EPS) with the dinitrosalicylic acid (DNS) method.

First, the tailspike activity spectrum was performed and compared with that of phage Stuart with spot tests. After overnight incubation at 37°C, all strains listed in Table 1 were spread into TSA plates and spotted with 10⁸ PFU/ml of phage Stuart or with 0.7 μg of the tailspike. After 8 h of incubation at 37°C, the results were visualized. Strains with clear spots surrounded with turbid haloes or just turbid haloes were scored as phage- or tailspike-susceptible, respectively. To assess the phage efficiency of plating (EOP) and exclude the possible effect of lysis from without (i.e., if phage is lysing the cells without infecting them), the suspensions of phage Stuart were serially diluted (10⁸ to 10³ PFU/ml) and spotted on the lawns of all strains listed in Table 1. EOP was calculated by dividing the titer of the phage (PFU/ml) for each strain by the titer obtained with the propagating host (PS#3). To assess the activity of tailspike, serial dilutions of the enzyme (7 to 0.0007 μg, equivalent to 10-μl drops of 10 to 0.001 μM) were prepared and spotted on the bacterial lawns.

Second, the tailspike EPS degradation activity was evaluated using EPS from phage Stuart host (PS#3) as described previously but with minor modifications (25). Overnight cultures grown in 2× concentrated TSB (100 ml) were pelleted (10,000 × g, 10 min) and suspended in 5 ml of 0.9% (wt/vol) NaCl. The suspension was further incubated with 5% phenol under agitation with a stir bar for 6 h. Cells were pelleted (10,000 × g, 10 min), and the supernatant was mixed with 5 volumes of 95% ethanol and kept overnight at −20°C to precipitate the EPS. After centrifugation (6,000 × g, 10 min), the precipitate was lyophilized and suspended in PBS. The EPS solution (5 mg/liter end concentration) was mixed with the tailspike at 0.07 μg/μl (equivalent to 1 μM end concentration) or PBS for 1 h at 37°C in the presence or absence of 5 mM EDTA. The reaction mixture was quenched by heating at 100°C for 5 min and centrifuged (8,000 × g, 2 min). An equal amount of dinitrosalicylic acid (DNS) reagent (Sigma-Aldrich) was added to the samples, the samples were heated at 100°C for 5 min, and the absorbance was measured at 535 nm to determine the concentration of reducing ends. Glucose was used as a standard.

The tailspike structure was characterized using (i) circular dichroism (CD) and (ii) gel electrophoretic analysis.

First, far-UV CD spectroscopy was used to analyze the protein secondary structure with a Jasco J-1500 CD spectrometer (61). The spectrum measurements were performed with the tailspike at 0.28 μg/μl (equivalent to 4 μM) in 10 mM potassium phosphate buffer (pH 7) using a wavelength range from 190 to 250 nm, with 1-nm steps, a scanning speed of 20 nm/min, high sensitivity, and a 16-s response time. Estimates of the tailspike secondary structures were made running the CD spectra in CONTINLL and set 4 optimized for the wavelength range of 190 to 240 nm at the DichroWeb server (62). We also employed thermal denaturation with 1°C/min increments to measure the secondary structure unfolding at 215 nm, from 30 to 100°C. The melting curves were fitted into a Boltzmann sigmoidal function.

Second, the tailspike oligomeric state was analyzed by comparing the protein molecular mass estimated based on the protein migration from a native and denaturing polyacrylamide gel electrophoresis.

The identification of the bacterial EPS as the phage Stuart receptor was made as previously described (25). Briefly, mid-exponential cultures of phage Stuart-insensitive (PS#1 and PS#2) and -sensitive (PS#3, PS#5, PS#11, and PS#15) strains at approximately 10⁸ CFU/ml were incubated with PBS as a control or with the tailspike at 0.07 μg/μl (equivalent to 1 μM end concentration) to remove the EPS. After 2 h of incubation at 37°C, cells were washed with TSB, incubated with the phage (at a multiplicity of infection [MOI] of 0.001) for 5 min, centrifuged, and titrated. The percentage of adsorbed phages was calculated by subtracting the number of nonadsorbed phages (supernatant) from the total amount of phages. Differences between the number of phages adsorbed to PBS- and tailspike-treated cells served to verify whether the EPS is the phage receptor.

The development of phage Stuart- and tailspike-insensitive bacterial variants was conducted similarly as previously described but with minor modifications (39). Overnight cultures of P. stuartii PS#3 were diluted 100-fold in TSB and incubated with PBS, phage Stuart (at an MOI of 1), or tailspike at 0.07 μg/μl (equivalent to 1 μM end concentration) for 72 h at 37°C with shaking. After each 24 h, cells were pelleted and washed twice in TSB prior to the addition of a fresh portion of PBS, phage Stuart, or tailspike, respectively. After 72 h of total incubation period, 20 colonies were isolated, purified three times on TSA plates, and tested for their sensitivity toward the phage Stuart and the tailspike using the spot test.

The serum killing assay was performed as previously described (63). Briefly, inocula of ≈10⁴ CFU/ml of phage Stuart-insensitive (PS#1 and PS#2) and -sensitive (PS#3, PS#5, PS#11, and PS15) strains were prepared. Serum was isolated from blood donated by healthy volunteers and mixed with (i) only cells or (ii) cells and the tailspike added simultaneously. Similar samples with heat-inactivated serum (at 56°C for 30 min) were used as controls. The serum was mixed with cells in a ratio of 1 volume of serum to 3 volumes of cells. The tailspike was added at 0.07 μg/μl (equivalent to 1 μM end concentration). After 1 h of incubation at 37°C, the CFU/ml were counted and serum killing activity was assessed by comparing their number to the number of CFU/ml in the initial inoculum.

The genomic sequence of the phage vB_PstP_Stuart genome was deposited in the NCBI GenBank database with the accession number MK387869.