K. pneumoniae KP/01, used as a host for bacteriophage infection, was isolated from a human clinical diabetic-foot sample from a male patient in May 2016 and identified by National Institute of Diabetes using the VITEK method for identification (Cairo, Egypt). Other clinical isolates of K. pneumoniae (n = 21), Proteus mirabilis (n = 18) and E. coli (n = 15) were also isolated by National Institute of Diabetes, for bacteriophage host-range analysis, from wound infection samples and provided to the microbiology research lab at Zewail City. Isolates were kept in tryptone soy broth (TSB; Oxoid, England) containing (w/v) 20% of glycerol, at −80°C. In the following experiments, bacterial strains were grown on tryptic soy agar (TSA; Oxoid, England) overnight, and isolated colonies of bacteria were grown at 37°C, in TSB, to reach OD₆₀₀ approximately 0.3.

PCR amplification was performed to confirm the identity of the K. pneumoniae isolate (KP/01) using specific primers for 16s RNA gene (forward primer: 5′-ATTTGAAGAGGTTGCAAACGAT-3′ and reverse primer: 5′-TTCACTCTGAAGTTTTCTTGTGTTC-3′; Woese and Fox, 1977; Woese et al., 1990). Thirty cycles were performed at denaturation temperature of 95°C for 30 s; annealing at 58°C for 60 s and extension at 72°C for 1 min looking for a PCR product of 133 bp length using an Applied Biosystems thermal cycler (Cady et al., 2012). The PCR product was run on a 1% (w/v) agarose gel to identify its size.

K. pneumoniae KP/01 strain was subjected to antibiotic resistance evaluation against a set of antibiotic discs including: tigecycline (TGC; 15 μg), imipenem (IPM; 10 μg), piperacillin-tazobactam (TZP; 100/10 μg), levofloxacin (LEV; 5 μg), linezolid (LZD; 30 μg), ceftazidime (CAZ; 30 μg), and cefepime (FEP; 30 μg) all from Oxoid (England). Antimicrobial sensitivity testing was performed for strains of K. pneumoniae, E. coli and P. mirabilisby using the disk diffusion methods in accordance with National Committee for Clinical Standards guidelines (Clinical and Laboratory Standards Institute, 1999). The antibiotics chosen are usually used for the treatment of diabetic foot infections in National Institute of Diabetes, due to their efficacy against members of the Enterobacteriaceae.

Bacteriophages were isolated from environmental water samples from freshwater in El- Maryoteyya-Haram area, Giza, Egypt. K. pneumoniae (KP/01) used as a bacterial host upon which the clear plaquing phage were selected for further characterization. The bacteriophage plaques were purified by repeated single plaque isolation using sterile micropipette tips (Adams, 1959). All isolated bacteriophages were amplified in liquid culture (TSB) and the lysates were centrifuged at 6,400 × g for 15 min at 4°C to remove remaining bacterial cells and debris (Marcó et al., 2012). The supernatant containing phages was then centrifuged for 1 h 15,300 × g at 4°C. The pellet was resuspended in SM buffer (100 mM MgSO₄.7 H₂O; 10 mM NaCl; 50 mM TrisHCl; pH 7.5) and filtered using 0.22 μm syringe filters (Chromtech, Taiwan). Bacteriophage titers were determined using double-agar overlay plaque assays (Mazzocco et al., 2009).

The morphology of bacteriophage ZCKP1 was investigated using transmission electron microscopy at the National Research Center (Cairo, Egypt). Formvar carbon coated copper grids (Pelco International) were immersed into phage suspension, the phage were fixed using glutaraldehyde (2.5% v/v), washed and stained using 2% phosphotungstic acid (pH 7.0). After drying, grids were examined using a transmission electron microscope (JEOL 1230).

DNA was prepared from bacteriophage ZCKP1 (10¹⁰ PFU/ml) to determine the genome size by pulsed field gel electrophoresis (PFGE; Senczek et al., 2000). Briefly, bacteriophage suspended in agarose plugs were digested with lysis buffer (0.2% w/v SDS [Sigma]; 1% w/v N-Lauryl sarcosine [Sigma]; 100 mM EDTA; 1 mg/ml Proteinase K [Fischer Scientific]), overnight at 55°C. Following washing 2 mm slices of agaraose containing DNA were inserted into the wells of a 1% w/v agarose gel. The gel was run by using a Bio-Rad CHEF DRII system, in 0.5 X Tris-borate-EDTA, for 18 h at 6 V/cm with a switch time of 30 to 60 s. The size of the genome was determined by comparison to standard concatenated lambda DNA markers (Sigma Aldrich, Gillingham, UK).

Genomic DNA was prepared from phage ZCKP1 (10¹⁰ PFU/ml) lysates by proteinase K treatment (100 μg/ml in 10 mM EDTA pH 8) followed by resin purification using the Wizard DNA kit (Promega, UK) following the manufacturer's instructions. DNA sequencing was performed using the Illumina MiSeq platform. The data consisted of 3.1 million paired-end sequence reads of 250 bp in length. Initial processing of the raw data and de novo assembly was performed using CLC Genomics Workbench version 11.0.1 (Qiagen, Aarhus, Denmark). ORFs were predicted from PHASTER and manually curated (Arndt et al., 2016). Nucleotide sequences appear under the GenBank accession number MH252123.

Using double-agar overlay plaque assays (Mazzocco et al., 2009), the lytic profile of phage ZCKP1 and other isolated phages was determined against a clinical isolate panel when spotted phage concentrations were not <10⁹ PFU/ml [34]. The experiment was performed using log phase bacteria. The panel included bacteria that cause osteomyelitis, including K. pneumoniae, P. mirabilis and E. coli. The lytic activity of bacteriophages was determined based on plaques of clear lysis. If ≥20 plaques were produced, the tested bacteria were regarded as being sensitive to the phages.

Bacteriophage ZCKP1 was tested in triplicate over eight decimal dilutions against all the susceptible bacterial strains lysed in the spot assays as previously described (Viazis et al., 2011). Conditions of these experiments were the same as spot test using log-phase bacteria. Thus, 200 μl of all bacterial isolates were added to top agar, and different dilutions of phages were spotted on petri dishes. The plates were incubated overnight at 37°C. Next day, EOP was estimated as the average PFU on target bacteria/average PFU on host bacteria.

The frequency of the emergence of bacteriophage insensitive mutants (BIMs) was estimated as previously described (O'Flynn et al., 2004). Phage ZCKP1 was mixed with bacterial host strains confirmed to be susceptible to the bacteriophage including strains of K. pneumoniae, P. mirabilis, and E. coli at an MOI of 100. After 10 min of incubation at 37°C, the suspension was serial diluted and spotted using double-agar overlay plaque assays. Plates were incubated overnight and BIM was calculated correspondingly by dividing bacterial viable counts remained after phage infection by initial viable counts. Experiments were conducted in triplicate.

One step growth curves were performed as previously described (Hyman and Abedon, 2009). Briefly, KP/01 strain was grown at concentration of 10⁸ and mixed with bacteriophage at multiplicity of infection of 1 and incubated at 37°C for 2 h. Directly after infection and every 10 min, aliquots of 200 μl were withdrawn and divided into two volumes of 100 μl. Chloroform was added to one of two volumes with a concentration of 1% (v/v); to set intracellular phages free while other 100 μl was left with no chloroform addition. After serial dilution, phage titer was estimated by spotting on top agar using double-layer method. Three replicates were conducted for each time interval.

The survival lysis characteristics of phage ZCKP1 were estimated KP/01 in the presence of ZCKP1 phage at multiplicities of infection of 0.1, 10 and 100 PFU/CFU was estimated in comparison to bacterial control at a temperature of 37°C (phage-free samples; Armon and Kott, 1993). Phage infective centers (IC) and plaque forming units (PFU) were also estimated, at different time intervals (0, 5, 10, 20, 30, 40, 60, 90, 120, and 180 min). IC is the amount of free phage particles released from the bacterial cells, without the need to add chloroform, while PFU refers to the number of nascent phage both inside and outside the bacterial cell. Briefly, two flasks were filled with either bacterial culture at a given concentration (control) or with bacterial culture at the same concentration and bacteriophage matching the desired MOI (Test). At every time interval, the concentration of bacterial control (B), bacterial survival (BS) IC, and PFU were simultaneously estimated. Bacterial concentration were determined using the Miles and Misra method (Miles et al., 1938), while phage concentration was estimated using double-agar overlay plaque assays by adding chloroform to the aliquot to be estimated in case of PFU determination, or not adding chloroform to calculate the IC.

Bacteriophage ZCKP1 was added to K. pneumoniae KP/01 in log-phase of growth, at 25°C, at an MOI of 1. Bacterial survival, number of infective centers, and number of plaque forming units were estimated periodically at different time intervals (0, 8, 24, 32, and 48 h).

The activity of ZCKP1 against established biofilms of KP1/01 was examined using a modification of previously described protocols (Cerca et al., 2005; Pettit et al., 2005). One hundred microliter aliquots of K. pneumoniae KP/01 (5 × 10⁶ CFU/ml) in 96-well flat-bottomed polystyrene microtitre plate (Sigma Aldrich) were incubated for 24 h at 37°C. Unattached planktonic cells were carefully removed. The number of bacterial cells in a biofilm per well were estimated to be 10⁷ CFU after 24 h (Mottola et al., 2013). Using different MOIs (5, 10, and 50), 100 μl aliquots of phage ZCKP1 diluted in TSB were added to each well, 1 day after biofilm establishment. Other wells received an equivalent amount of TSB as positive controls. In a parallel experiment, phage was introduced to wells every 4 h carefully replacing the previous suspension (containing TSB, planktonic cells and released phages) without disturbing the established biofilms. The biomass of preformed biofilms was quantified by staining with crystal violet (0.2% w/v). Following washing to remove excess dye with PBS, the crystal violet was solubilized in ethanol (95%). The absorbance was measured using a microplate reader at OD₆₀₀ (Biotek, USA). The bacterial counts in biofilms were estimated using an MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay (Serva Electrophores, Germany) as described by Cady et al. (2012). The absorbance was then measured at 570 nm at 4, 12, and 24 h, using a microplate reader (BioTek, USA). Control and test samples were assayed in triplicate.

The temperature stability of phage ZCKP1 (10¹⁰ PFU/ml) was evaluated at 45, 55, 65, 75, 85, and 95°C, at 10 min intervals, over 1 h in adjusted water bath incubator. Immediately after incubation, serial dilutions of phage were spotted in triplicate, using standard double layer technique; on a lawn of host strain (KP/01) to estimate phage titers as previously described (Capra et al., 2004; Hammerl et al., 2014).

The bacterial counts of ZCKP1 at different pH values (5, 6, 7, 8, and 9) was determined after 1 h incubation, followed by determining the phage titer as previously described (Hammerl et al., 2014). Different pH values were achieved in SM phage buffer to maintain comparative conditions.

In all data sets, test and control sets were compared using Student's t-test. A significance level of 0.05 was applied in all cases. Analytical statistics were undertaken using GraphPad PRISM version 7.00 for Windows (GraphPad Software, La Jolla, USA).

The identity of the KP/01 strain was confirmed to be K. pneumoniae by PCR, by the presence of 133 bp band corresponding to conserved region in 16s RNA gene of K. pneumoniae, following amplification with the specific primers. The antibiotic sensitivity of K. pneumoniae isolate KP/01 was tested using the disc diffusion method and the results showed that K. pneumoniae isolate KP/01 was sensitive to tigecycline (TGC), imipenem (IPM) and piperacillin-tazobactam (TZP) but resistant to levofloxacin (LEV), linezolid (LZD), ceftazidime (CAZ) and cefepime (FEP).

Bacteriophages were isolated from freshwater near the pyramids of Egypt in Giza. Selection of the bacteriophage was undertaken upon serial passage according to their ability to lyse a broad range of K. pneumoniae isolates and other pathogens causing osteomyelitis, generate reproducible clear zones of lysis, produce hallow zones around lysis zones indicative of exopolysaccharide depolymerase activity and capable of replication to produce high titers on the selected host with respect to time. Bacteriophage ZCKP1 fulfilled these criteria.

Electron microscopy revealed that ZCKP1 had an icosahedral head and contractile tail with collar, and base plate, and therefore typical of phages belonging to the family of Myoviridae (Figure 1). The proportions of the phage head and tail length were also typical of the Myoviridae with the head size being 80 ± 0.7 nm while tail length was calculated to be 138.5 ± 2.5 nm.

Bacteriophage ZCKP1 contains a double-stranded DNA genome estimated to be 160 kbp by PFGE, which is comparable to values indicated by International Committee on Taxonomy of Viruses (ICTV) for bacteriophages belonging to the Myoviridae family. DNA sequencing of the phage DNA enabled de novo assembly and accurate size determination of a circular permuted genome of 150,925 bp with a G + C content of 39.1%. The genome contained 267 open reading frames, the majority of which are hypothetical proteins or recognized in BLASTP database searches as phage proteins without any ascribed function. Reading frames for which putative functional information could be ascribed to the products appear in Supplementary Table 1. Notably these include the phage structural proteins, nucleotide metabolism and components of the replication machinery that are conserved amongst Myoviridae infecting hosts within the Enterobacteriaceae. Of interest are enzymes that have the potential to modify infected cell surface polysaccharides that may impede superinfection. These include an O-antigen biosynthesis protein, a glycosyltransferase and a wcaM superfamily protein associated with colonic acid biosynthesis clusters present in Enterobacteriaceae that feature exopolysaccharide production. Four genes encoding proteins related to tellurite resistance are present. Tellurite resistance is frequently used for selection in culture isolation media but is not used for antimicrobial therapy. The genes are thought to contribute to colicin and phage resistance (Taylor and Summers, 1979), which may provide reasons for their presence in phage ZCKP1 in that colicin resistance will provide a selective advantage to the phage infected cell and phage resistance to prevent superinfection. Also of note the phage encodes a member of the hydrolase 2 superfamily implicated in bacterial cell wall hydrolysis. The nearest database phage sequence was PHAGE_Escher_phAPEC8 that infects avian pathogenic E. coli and is also a member of the Myoviridae (Tsonos et al., 2012).

The host range of five different phages isolated from freshwater, including phage ZCKP1 were tested on bacteria that were isolated from diabetic patients suffering from osteomyelitis. The ZCKP1 phage was capable of producing lysis zones (≥20 plaques) on 15 out of 21 K. pneumoniae isolates, 5 out of 18 P. mirabilis isolates and 9 out of 30 E. coli isolates, while other phages did not display a comparable spectrum of activity against the K. pneumoniae isolates (Table 1). A range of EOP for ZCKP1 phage was observed against different species of Enterobacteriacae (Supplementary Table 2). For K. pneumoniae seven phages demonstrated EOPs similar to the multidrug resistant host strain. For P. mirabilis, all susceptible strains showed EOP <0.1, whereas for E. coli six strains supported replication with EOPs approaching that of the permissive K. pneumoniae hosts (Table 2).

BIMs were recovered following high multiplicity infections (100) of host bacteria K. pneumoniae, P. mirabilis and E. coli with bacteriophage ZCKP1 at 37°C. Mutational frequencies of 7.5 × 10⁻⁵ ± 1.7 × 10⁻⁴ and 3.7 × 10⁻⁵ ± 6.8 × 10⁻⁵ were determined for Klebsiella and E. coli, respectively where K. pneumoniae KP1 alone exhibited a lower frequency of 5 × 10⁻⁶ ± 4.04 × 10⁻⁶.

A single-step growth curved demonstrated bacteriophage virions were naturally released from bacterial cells after 30 min: the latent period which is the time taken for phages to be assembled and released after infection. However, viruses were assembled 10 min before. This was indicated by eclipse period that was estimated to be 20 min, as chloroform aids new phage particles to free from bacterial cell wall (Figure 2). Burst size was estimated to be ~110 virions per single bacterium.

The infection and lysis characteristics of phage ZCKP1 were estimated at different MOIs, over a period of 3 h (Figures 3A–C) in a growing culture of K. pneumoniae KP/01 (Figures 3A–C). K. pneumoniae KP/01 was lysed by phage ZCKP1 at each MOI tested but the MOI of 100 reduced the viable bacteria from 9.0 log₁₀ CFU/ml to below the limit of detection at 37°C by 2 h (Figures 3A–C). Under these circumstances the reductions in bacterial count were not accompanied by a measurable rise in phage titer (Figure 3C). Phage replication was observed at lower MOI, which coincided with the commencement of the fall in viable count.

A single application of ZCKP1 to established biofilms of K. pneumoniae KP/01 resulted in a reduction crystal violet stainable biofilm content (P < 0.01; Figure 4A) and the percentage of viable cells observed by MTT staining (P < 0.01; Figure 4C) after 4 h. The most effective treat represented the highest MOI (50 PFU/CFU). However, following this disruption there was recovery in biofilm estimates accompanied by a recovery in cell viability. Multiple treatments of phage ZCKP1 on established K. pneumoniae KP/01 biofilms at 4 h intervals resulted in significant reductions in biofilm content and prevented the recovery of cell viability throughout the 24 h period of the experiment (P < 0.01; Figures 4B,D).

The stability of phage ZCKP1 at different temperatures and pH values was investigated (Figures 5A,B). Phage titers were stable, at approximately 10⁹ PFU/ml, for 1 h at temperatures of 45 and 55°C. The phage titer decreased after 40 min at 65°C to 10⁸ PFU/ml, and continued to decline below 10⁷ PFU/ml after 1 h. A significant decline (P < 0.005) was observed when phages were incubated at 75 and 85°C. However, phage could still be recovered after 1 h at 75°C at a titer of 10³ PFU/ml. Phage could not be recovered after 40 min at 85°C. Acidic pH of < 6 significantly (P < 0.005) reduced the phage stability after 1 h. The optimum stability was observed to be pH 6 but persisted at alkaline pH values to pH 9 (Figure 5B).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.