Thirty-two isolates of K. pneumoniae were subcultured from biobanked clinical isolates, routinely collected from patients in tertiary care hospitals in Cairo, Egypt, between 2017 and 2019. The isolates were collected from sputum (n = 20) and endotracheal aspirate (n = 12).

The biobanked isolates had been identified by the Vitek 2 automated system (bioMérieux, Marcy l’Étoile, France). Before further subcultures, the bacterial identity of each isolate was verified and confirmed by Gram-staining, followed by culture on differential selective media (MacConkey agar, Neogen, UK) and typical biochemical reactions for Enterobacteriaceae (Farmer et al., 1985).

The identity of the bacterial strain used for phage propagation was further confirmed by the polymerase chain reaction (PCR) as previously described (Turton et al., 2010), followed by sequencing of its 16S–23S rRNA internal transcribed spacer. Conserved, predesigned primers (Woese and Fox, 1977) for that region (FP: 5’-ATTTGAAGAGGTTGCAAACGAT-3’ and RP: 5’-TTCACTCTGAAGTTTTCTTGTGTTC-3’) were synthesized by LGC Biosearch Technologies, UK. After PCR, the amplicon size was confirmed by gel electrophoresis, and the DNA was extracted from the 1% (w/v) agarose gel, and sequenced at Macrogen, Korea. The obtained sequence was checked by BLASTn against the rRNA/ITS database at the National Center for Biotechnology Information (NCBI) portal.

For phage isolation purposes, bacterial strains were cultured on tryptic soy broth (TSB, Oxoid, UK) media at 37°C in a shaking incubator at 150 rpm. Also, TSB supplemented with 20% glycerol was used as a maintenance medium to preserve all isolates at -80°C.

The antimicrobial susceptibility of the bacterial isolates was estimated according to the Clinical and Laboratory Standards Institute 2019 recommendations (CLSI) (CLSI, 2019). Colonies of overnight culture were aseptically suspended in sterile isotonic saline solution to be visually comparable to a 0.5 McFarland standard. The bacterial suspensions were swabbed aseptically and uniformly on Mueller-Hinton agar (MHA) (Oxoid, UK). The antimicrobial discs were aseptically added with an antimicrobial susceptibility disc dispenser (Oxoid, UK). The cultures were incubated at 37°C for 18 h. The diameters of the inhibited zone around the discs were measured in triplicates, and the mean diameters were calculated and interpreted in reference to the breakpoints of CLSI 2019 (CLSI, 2019) into either susceptible, intermediate or resistant. The 18 tested antimicrobials (belonging to 10 different antimicrobial classes) were piperacillin, piperacillin-tazobactam, cefuroxime, cefepime, doripenem, meropenem, ertapenem, imipenem, aztreonam, azithromycin, tobramycin, kanamycin, gentamicin, amikacin, ciprofloxacin, levofloxacin, trimethoprim-sulfamethoxazole, and doxycycline. The antimicrobial discs were purchased from Oxoid, UK ( Table S1 ).

The results of the antimicrobial susceptibility assay of all isolates were further analyzed to calculate their multiple antibiotic resistance (MAR) index (Krumperman, 1983). In addition, we developed a modified MAR index that considers intermediately susceptible isolates. The modified MAR index is the ratio of antimicrobial agents to which a bacterial isolate is not susceptible (intermediately susceptible or resistant) to the total number of assayed antimicrobial agents. While the MAR index assigns a score of 1 to any antibiotic to which the bacterial isolate is resistant and zero to other cases, the modified MAR index scores resistance as 1 and intermediate susceptibility as 0.5. As Kahlmeter et al. (2019) recommended, resistant and intermediate isolates were considered ‘not susceptible,’ a term they suggested instead of ‘nonsusceptible’ (Kahlmeter et al., 2019). Not susceptible bacterial isolates with one or more antimicrobial agents in three or more classes were considered MDR (Magiorakos et al., 2012).

The bacterial isolates used to determine the phage host range were screened for virulence genes by PCR. These genes were classified into siderophores (entB, iutA, and irp2), types 1 and 3 fimbriae (fimH, mrkD), located in either the core or accessory genome.

The sequences of the targeted virulence genes were retrieved from the NCBI database. Primers for the virulence genes were designed by Primer-BLAST (Ye et al., 2012) and analyzed by OligoAnalyzer (URL://https://eu.idtdna.com/pages/tools/oligoanalyzer). The primers were tested by the in silico PCR amplification tools (Bikandi et al., 2004), and BLASTn checked the obtained sequence of PCR products. The primers were synthesized by Willowfort, UK ( Table S2 ).

For virulence gene amplification, bacterial colonies were boiled in sterile saline solution, and a colony PCR approach was used (Gussow and Clackson, 1989; Azevedo et al., 2017). The PCR products were separated by agarose gel electrophoresis (Sigma Aldrich, Germany), and 100 bp DNA Ladder RTU (GeneDireX, India) was used.

Several liquid sewage samples were collected from urban and medical sources in different Egyptian cities. The crude sewage samples were used without any treatment. Each isolate of K. pneumoniae was cultured overnight in TSB and incubated at 37°C with 150 rpm shaking. A mixed culture of all K. pneumoniae isolates was prepared by transferring 0.5 ml of each culture into a 100 ml conical flask with fresh 50 ml TSB. The mixed culture was incubated overnight at 37°C and 150 rpm.

For enrichment, each collected sewage sample was mixed with an equal volume of the mixed culture of K. pneumoniae isolates in a sterile 15 ml centrifuge tube (Carlson, 2005; Twest and Kropinski, 2009). Each enriched culture was incubated for 4 h at 37°C in a static incubator. Next, the culture was centrifuged at 2,300 × g for 20 min, and 1 ml of the supernatant (phage lysate) was aseptically collected and stored at 4°C. For culturable phage screening, 10 μl of each lysate was spotted twice on tryptic soy agar (TSA, DifcoTM, BD, Franklin Lakes, NJ, USA) plates supplemented with 1% (w/v) top soft TSA, which was inoculated with 100 µl of bacterial broth cultured in its exponential growth phase. The plates were incubated overnight at 37°C. On the following day, the formed phage plaques on the bacterial lawns were excised and suspended in 500 µl sodium magnesium (SM) buffer [5.8 g of NaCl; Fisher Chemicals: S/3120/60, 2 g of MgSO₄.7H₂O; Fisher Chemicals: M/1050/60, 5 ml of 2% (w/v) Gelatin; Sigma: G-1393 and 50 ml of 1 M Tris-HCl; Sigma: T-2663 at pH 7.5, in a liter of distilled water (dH₂O)]. The buffer was initially sterilized in an autoclave and by filtration through a 0.22 μm cellulose acetate (CA) syringe filter (Standard Membrane Filtration Limited, China), and then was left for 2 h at 4°C for phage particle elution. The SM buffer for each plaque was aseptically transferred into new microcentrifuge tubes, and then centrifuged at 2,300 × g and 4°C for 20 min. The supernatant was transferred into a new microcentrifuge tube and stored at 4°C. Eventually, the isolated phages were further spotted on their bacterial K. pneumonaie hosts for confirmation.

The isolated bacteriophages were serially diluted 10 times in SM buffer, then spotted on the surface of dried bacterial lawns and incubated aerobically at 37°C following the spot test (Carlson, 2005). The spotted plates were examined for bacteriophage plaques after 24 h. Single plaques were picked from the top agar layer using a sterile micropipette tip. For phage particle elution, the excised plaques were suspended in a 300 μl SM buffer, vortexed for 10 sec, then left for ≥ 24 h at 4°C. Each phage lysate was centrifuged at 5,500 × g for 20 min, and the supernatant was collected aseptically into new sterile microcentrifuge tubes, which were kept at 4°C. This purification procedure was repeated six times for each isolated bacteriophage to produce purified phage stocks. Subsequently, aliquots of purified bacteriophages were used for propagation.

The purified bacteriophages were propagated by the plate lysate method, according to Bonilla et al. (2016), with some modifications. Ten microliters of the purified lysate and 100 µl of a fresh culture of K. pneumoniae host strain were mixed with 5 ml melted 1% (w/v) soft TSA, which were poured on a plate containing TSA base agar. Thirty plates were prepared as above, incubated overnight at 37°C, and plaques were aseptically scraped in 3 ml of sterile SM buffer, and left at room temperature for 15 min. The buffer was collected from all plates in one tube and centrifuged at 5,500 × g at 4°C for 20 min. The supernatant of phage lysate was transferred to a new sterile tube. Chloroform (1% v/v) was added to the phage lysate and left for 2 min at 4°C; then, the upper layer was extracted into a new sterile tube. The phage lysate was further centrifuged at 5,500 × g for 20 min, and the collected supernatant was filtered through 0.22 μm CA syringe filters. The phage preparation was concentrated by centrifugation at maximum speed (18,000 × g) for 90 min at 4°C, followed by pellet resuspension in 3 ml SM buffer (modified from Carlson, 2005). Plaques corresponding to the concentrated phage lysates were enumerated by spot test (Carlson, 2005).

To determine the phage host range, we used the previously described spot testing assay (Kutter, 2009) on 32 bacterial isolates of K. pneumoniae ( Figure 1 ). The purified phage lysates were spotted twice on bacterial lawns seeded in TSA plates (TSB with 1% top agar) with 100 μl of each test strain. After overnight incubation, the plates were checked, and their plaques were scored according to their clarity (Kutter, 2009). A clear plaque was scored +3, indicating complete lysis, while an opaque plaque with a hazy or turbid background was scored +2. Spots with non-confluent plaques were scored +1, and spots that showed no plaques were scored 0. For confirmation, the spot testing assay was repeated for bacterial hosts with plaques of scores > +2.

Phage effectiveness against K. pneumoniae host strains was further assessed by the efficiency-of-plating method (EOP) as previously described (Kutter, 2009). Briefly, a stock of the phage was serially diluted 10-fold (10¹ to 10⁸). Ten microliters of each dilution were spotted in triplicates on a fresh lawn of each susceptible bacterial isolate identified according to the host range assay. After overnight incubation, the plaques of the same phage titer were enumerated, and the average plaque-forming unit (PFU) count for spot triplicate was calculated for each bacterial isolate. The relative EOP was calculated as the ratio of the average PFU count of the same phage dilution on each bacterial host to the maximum PFU count observed (Kutter, 2009). The EOP was classified into four levels of efficiency from ‘high production’ to ‘no production’ ( Table S3 ) (Mirzaei and Nilsson, 2015).

Purified high-titer lysate of 10¹⁰ PFU/ml of the phage was prepared and sent to the TEM core facility at the National Research Center (NRC), Cairo, Egypt. The prepared lysate was added to a copper grid, fixed with 2.5% (v/v) glutaraldehyde and negatively stained by 2% (w/v) phosphotungstic acid (pH 7.2). Micrographs were obtained by a JEOL 1230 transmission electron microscope (NRC, Cairo, Egypt), and the dimensions of the virion were measured by ImageJ version 1.53n with reference to the scale bar of 100 nm that was generated from the microscope.

The bacteriolytic activity of the phage was evaluated in vitro against its isolating bacterial host at different multiplicity of infection (MOI) values (Carlson, 2005). The phage lysate concentrations 10⁶, 10⁷, 10⁸, and 10⁹ PFU/ml were prepared to achieve MOIs 0.1, 1, 10, and 100, respectively. The challenged bacteria were freshly grown in TSB medium to mid-log phase with a concentration of 10⁶ - 10⁷ colony-forming units (CFU)/ml. The concentration of bacterial suspension was determined according to its optical density at 600 nm. For each MOI, 20 ml of mid-log bacterial suspension was prepared, then divided into tubes labeled as T and C (for test and control, respectively), each containing 10 ml. The added volume of the phage lysate was calculated according to each MOI.

The two tubes were incubated at 37°C with shaking at 100 rpm. The bacteriolytic activity of the phage was evaluated through 11 time points (0, 10, 20, 30, 45, 60, 75, 90, 120, 150, and 210) over 3.5 h of incubation. At each time point, a sample of 100 µl was drawn from both tubes (T and C). The drawn samples were serially diluted 10-fold from 10^-1 to 10^-10 in TSB, and each dilution was spotted on agar plates. The titer of the uninfected bacteria was determined as CFU/ml from plates of tube C (Fayez et al., 2021). Additionally, the titer of surviving bacteria after infection and the titer of free phage as PFU/ml were determined from plates corresponding to tube T. The bacterial titers were assayed by the plate count method (Lee, 2008), while the phage titers were determined by spot test (Carlson, 2005). The titers of surviving bacteria in the assayed MOIs were determined to evaluate the most efficient bacteriolytic MOI (Liu et al., 2020).

Moreover, a portion of phage-treated bacterial suspension of MOI 0.1 was collected at every time point. The collected samples were divided into two tubes, one untreated with chloroform for infective centers quantification and the other was treated with 1% (v/v) chloroform to induce lysis and burst of the intact bacterial cells to free intracellular phages. Next, the samples were diluted and spotted on bacterial lawns for post-burst phage titer determination. The burst size was calculated as the ratio of the titer of released virions at plateau to the initial virions titer (Carlson, 2005). These assays were conducted in triplicates.

The phage stability was assessed under various conditions of temperature and pH. Different temperatures were used to cover both storage and different thermal conditions (- 20°C, 4°C, 40°C, 60°C, 70°C, and 80°C). Phage aliquots were adjusted at 10¹⁰ PFU/ml as starting titers for each temperature. The phage titer of each aliquot were determined at indicated time points for 2 h (10, 20, 30, 60, and 120 min) and spotted after 24 h for an extended time point.

Moreover, the phage stability was monitored over a broad pH spectrum ranging from 2 to 13. The SM buffer pH values were adjusted by either 1M hydrogen chloride (HCl) or 1M sodium hydroxide (NaOH). The phage titers for each pH condition were determined after 1 hour and 24 h of incubation at 4°C. All phage titers were determined by spot test (Carlson, 2005). The temperature and pH stability experiments were conducted in triplicates.

The anti-biofilm activity of the phage was evaluated at different MOIs (100, 10, 1, 0.1, 0.01, 0.001, and 0.0001) for two different phenotypes: (i) inhibition of biofilm formation and (ii) clearance of formed biofilm. The biofilm formation, staining, and measurement were conducted by the microtiter plate biofilm assay (Merritt et al., 2011) with minor modifications.

A fresh TSB-diluted bacterial culture was grown to a concentration of 10⁵ CFU/ml (exponential phase), estimated by optical density at 600 nm. A volume of 125 μl of the culture was added to wells of flat-bottomed polystyrene microtiter plates (Greiner Bio-One, Portugal). In each microtiter plate, a row of 6 wells containing untreated culture was used as a negative control.

The 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was used to evaluate the effect of phage vB_Kpn_ZCKp20p on the viability of normal human skin fibroblast (HSF) cells. The HSF cells (passage 12) were grown in high glucose Dulbecco’s Modified Eagle’s Medium (DMEM, Biowest, France) supplemented with 10% fetal bovine serum, 100 units/mL penicillin, and 100 mg/mL streptomycin. The cells were seeded with density 8 x 10³ cells/well in 96-well cell culture plates (CELLSTAR, Greiner Bio-One, Portugal) and grown in a carbon dioxide incubator at 37°C.

A purified lysate of phage vB_Kpn_ZCKp20p in SM buffer was 10-fold serially diluted in DMEM media to be assayed at different concentrations (10⁹,10⁸,10⁷, and 10⁶ PFU/ml) on a proliferative monolayer of fibroblasts in the wells. The phage effect on the cells’ growth was evaluated after 24 h and 48 h. A negative control of untreated cells in DMEM media and a vehicle control of SM buffer diluted in DMEM media with untreated cells were prepared.

After incubation at each time point, the used DMEM media was replaced with fresh 100 µl DMEM containing 10% MTT labeling reagent (final concentration 0.5 mg/ml) and left in the incubator for 4 h. Then, the media was carefully withdrawn, and an equal volume of DMSO was added per well to solubilize the formed crystals of formazan for 20 min. Next, the wells’ absorbance was measured at 570 nm in a FLUOstar Omega Microplate Reader. The OD values were used to calculate the viability of the phage-treated cells as percentage of the untreated cells (negative control) by the following formula (Carmichael et al., 1987; van Meerloo et al., 2011; Rahimzadeh Torabi et al., 2021; Rezk et al., 2022). The cell viability assay was conducted in four technical replicates and three biological replicates for each incubation time point.

The genomic DNA of the phage was extracted from 10 ml of purified high-titer (10¹⁰ PFU/ml) phage lysate by the phenol-chloroform–isoamyl alcohol method, as detailed in Sambrook et al. (2001), with some modifications. Initially, the phage capsid was lysed to release the contained DNA by incubation for an hour at 56°C with proteinase K and sodium dodecyl sulfate (SDS) (10% w/v), which were added to a final concentration of 50 ug/ml and 0.5%, respectively. The solution was then thoroughly mixed with an equal volume of phenol: chloroform: isoamyl alcohol (25:24:1) to separate genomic DNA from proteins. The mixture was separated into two phases by centrifugation at 18,000 × g for 10 min, an aqueous phase on the top from which DNA was collected and an organic phase containing proteins, which was discarded.

The extracted DNA in the aqueous phase was precipitated overnight at -20°C by 1:10 volume of 3M sodium acetate (pH 5.2) and 2:1 volume of ice-cold isopropanol. The precipitated DNA was pelleted at 18,000 × g for 10 min by centrifugation. The supernatant was discarded, and then the pellet was resuspended in 90% ice-cold ethanol and transferred into a new 1.5 ml tube. The extracted DNA pellet was washed twice in 70% ethanol, left to dry, and finally resuspended in 100 μl nuclease-free water. The DNA concentration and quality were measured by FLUOstar Omega Microplate reader (BMG LABTECH, Germany).

The extracted DNA was sequenced at the core genomics facility of Zewail City of Science and Technology, with the standard Illumina protocols. Briefly, a genomic library of the extracted DNA of the phage was constructed by the Illumina Nextera tagmentation protocol (Illumina, Cambridge, UK). An Illumina v3 sequence cassette was used for sequencing the tagmented library for 600 cycles on an Illumina MiSeq platform.

Sequence reads were quality checked by FastQC (v0.11.9) and de novo assembled by Unicycler (v0.4.8) on the PATRIC platform (v3.6.12) (Davis et al., 2020), currently a part of the BV-BRC portal (https://www.bv-brc.org/). In the PATRIC assembly pipeline, the assembled genome was polished by Pilon (v1.23), and the resulted contig was plotted by Bandage (v0.8.1). The assembly statistics were performed by QUAST (v5.0.2) and SAMtools (v1.3).

Next, the single assembled contig was annotated with the Rapid Annotation using Subsystem Technology Toolkit (RASTtk) pipeline (Aziz et al., 2008). The annotation scheme was customized to start with “annotate-proteins-phage” followed by “annotate-proteins-kmer-v2”. tRNAscan-SE (v2.0) was used to identify tRNA genes (Chan and Lowe, 2019).

After RASTtk annotation, a second-round of annotation was performed for confirmation of assigned functions, or for assigning functions to proteins with unassigned functions. Several tools were used for that purpose, including NCBI BLASTp (against the non-redundant protein database), HHPred, InterPro Scan, and UniProt. Accordingly, the annotated proteins were manually edited in the UGENE software, v43.0 (Okonechnikov et al., 2012), and a final GBK file was generated. A genomic map of the phage was generated by Proksee web-based tool (Jason R. Grant et al., 2022) by CGView family tools (Stothard et al., 2019).

The phage lifestyle was predicted by BACPHLIP (Galaxy Version 1.0) (Hockenberry and Wilke, 2021), which predicts lifestyle based on conserved domains of integrases, excisionases, recombinases, transposases, and partitioning proteins (ParA and ParB). Moreover, the phage therapeutic suitability was evaluated by PhageLeads, which analyses the phage genome for temperate genetic markers, AMR, and virulence genes. Other tools were used to further test the feasibility of the phage for therapy, including RGI Resistance Gene Identifier (RGI v5.2.1, CARD v3.2.3 (Alcock et al., 2020), DBETH (Chakraborty et al., 2012), and VRprofile2 (Zankari et al., 2017; Bortolaia et al., 2020; Wang et al., 2022). PhageTerm (Galaxy v1.0.11) was used to predict the genome termini and packaging from sequence reads (Garneau et al., 2017).

The topology of the phage predicted proteome was analyzed for transmembrane domains by DeepTMHMM, which uses a deep learning protein language model-based algorithm (Hallgren et al., 2022). The genomic sequence was scanned by Phage Depolymerase Finder (PhageDPO) (Galaxy Version 0.1.0) prediction tool to find genes with putative depolymerase function. This tool uses two models, the Support Vector Machine (SVM) and the Artificial Neural Network (ANN) models, to analyze the genomic sequence (Duarte, 2021).

The whole genome sequence of the phage was checked for similarity with deposited sequences in the GenBank-NCBI database by BLASTn. The top hits with the highest scores and identity ≥ 90% were compared to the phage genome, and the average nucleotide identity (ANI) based on BLAST+ (ANIb) was also calculated by JSpeciesWS (Richter et al., 2016). The genomes with the highest ANI were compared to the phage genome and visualized by ProgressiveMauve (Darling et al., 2010). Additionally, these phages were compared in terms of the proteome by the Proteome Comparison tool of the PATRIC platform, which performs protein bidirectional sequence-based comparison by BLASTp (Overbeek et al., 2014).

Genome-based taxonomy was conducted according to the guidelines by Turner et al. (2021). We applied a proteome-based clustering strategy using the ViPTree server (Nishimura et al., 2017) for family-level classification. The analysis was conducted with the closely related phages and other dsDNA prokaryotic viruses. Phages with highest VipTree tBLASTx scores (S_G ) were selected to construct a more detailed rectangular proteomic tree.

For further classification to taxonomic levels below the family level (genus and species), a phylogenetic analysis of genome-genome distance of the isolated phage with BLASTn top hits, phages of highest (S_G ) on ViPTree and an outgroup of other morphotypes and families was performed. The genome-genome distance was computed by Virus Classification and Tree Building Online Resource (VICTOR) (Meier-Kolthoff and Göker, 2017) based on Genome-BLAST Distance Phylogeny (GBDP) method (Meier-Kolthoff et al., 2013). As indicated in the VICTOR documentation, minimum evolution trees were inferred from the calculated intergenomic distances, and FASTME was used for branch support computation. The analyses were supported by 100 pseudo-bootstrap replicates. Trees were rooted at the midpoint (Farris, 1972) and visualized by ggtree (Yu, 2020). The OPTSIL program was used to estimate taxon boundaries at different taxonomic levels (Goker et al., 2009) at recommended clustering thresholds (Meier-Kolthoff and Göker, 2017) and an F value (fraction of links required for cluster fusion) of 0.5 (Meier-Kolthoff et al., 2014). Moreover, VIRIDIC (Virus Intergenomic Distance Calculator) was used to identify intergenomic similarities of the isolated phage genome to the closest phage homologs in BLASTn and other Caudoviricetes phages that were filtered from the NCBI virus database. The filtered phages from the NCBI were select according to the following criteria, Caudoviricetes, complete RefSeq genome, genome length between 40Kb - 60Kb, and their host type is Gammaproteobacteria. The FASTA sequence of the filtered phages were downloaded and used for VIRIDIC analysis. VIRIDIC applies the traditional algorithm to calculate pairwise intergenomic distances/similarities according to ICTV recommendation (Moraru et al., 2020).

Additionally, genomic relationships between the phage and reference phages genomic sequences on NCBI-GenBank were analyzed with the alignment-free method, PhageClouds (Rangel-Pineros et al., 2021), which calculates intergenomic distances with dashing based on a threshold of 0.25.

Pan-genomic analysis was performed by CoreGenes5 to find shared signature genes of the phage and its close relatives (Davis et al., 2022). The amino acid sequence alignments of the highly conserved signature genes, encoding the terminase large subunit, the major capsid protein, and the DNA polymerase enzyme, were used for further confirmation of genus- and family-level phylogeny. The conserved gene-based phylogenetic analysis was conducted for the isolated phage with its close relatives against an outgroup of podoviruses and myoviruses. The amino acid sequences of the selected conserved genes of the selected phages were aligned by ClustalW, and the evolutionary history was inferred by the maximum likelihood method with Le_Gascuel model (Le and Gascuel, 2008) for terminase large subunit, major capsid protein, and JTT matrix-based model (Jones et al., 1992) for DNA polymerase III beta subunit. The analysis was supported with 1000 bootstrap replicates. The evolutionary rate differences among sites were modeled by a discrete gamma distribution. Tree branch lengths were measured in number of substitutions per site, while all gaps and missing data were entirely removed. The trees of the phylogenetic analyses were constructed in MEGA (v11) (Tamura et al., 2021).

All statistical analyses were performed by GraphPad Prism 9.1.1 for Windows (GraphPad Software, San Diego, California, USA). Data are presented as means and standard deviation of the mean (± SD). Control and test sets were compared by t-test/one-way ANOVA at a significance level of P< 0.05.

The genomic sequence of phage vB_Kpn_ZCKp20p was deposited in the NCBI nucleotide database (GenBank) under accession number OP373729.

Six distinct phages were isolated from urban and medical sewage of three different cities to ensure source diversity. Phage vB_Kpn_ZCKp20p, the focus of this study, was isolated from the medical sewage of a healthcare facility in Cairo.

The host range broadness and antimicrobial profile distinguished the isolated phages as potential therapeutic agents. Out of 32 tested Klebsiella isolates, the isolated phages lysed 9 to 12 hosts: Phages ZCKp24r2, ZCK4pz1, and vB_Kpn_ZCKp20p formed clear plaques with all affected hosts, while phages ZCKp4t, ZCKp4q, and ZCKp16r1 formed clear plaques on only 90%, 77%, and 50% of their host lawns, respectively ( Figure 1D ).

When bacterial isolates were arranged according to their modified-MAR indices from the highest (0.92) to the lowest (0.36), a negative correlation between the host ranges of the isolated phages and bacterial resistance profile was observed (Spearman’s correlation r_s = -0.6139, P = 0.00019, Figure 2 ). No bacterial host with a modified-MAR index above 0.69 was susceptible to any of the isolated phages, and the most susceptible bacterial isolate, to infection by both phages ZCKp4q and vB_Kpn_ZCKp20p, was K2 (modified-MAR index = 0.69). In contrast, the bacterial hosts with modified-MAR indices ≤ 0.5 were susceptible to four to six phages, except for one isolate (K17).

These preliminary results reflected that phage vB_Kpn_ZCKp20p had the broadest host range, of which the bacterial isolates had the highest average of modified-MAR indices; thus, phage vB_Kpn_ZCKp20p was selected for further characterization and genomic analysis.

The pathogenicity of K. pneumoniae is associated with genes encoding for virulence factors; these genes are either part of the core or accessory genome. In this study, the susceptible bacterial isolates to phage vB_Kpn_ZCKp20p were screened by PCR for siderophore-related genetic determinants, including enterobactin (entB), aerobactin (iutA), and yersiniabactin (irp2), in addition to, type 1 fimbriae (fimH). The entB and iutA genes, associated with enterobactin biosynthesis and aerobactin uptake, respectively, were detected in all tested isolates. Additionally, the yersiniabactin genetic determinant was identified in all isolates, except K22. Lastly, fimH and mrkD were identified in all bacterial isolates susceptible to phage vB_Kpn_ZCKp20p.

Given the clinical importance of biofilm formation (as they boost bacterial resistance and virulence), the antibiofilm activity of phage vB_Kpn_ZCKp20p was experimentally assessed at a wide range of MOIs. Both biofilm formation inhibition and disruption of a pre-formed mature biofilm were evaluated. At all assayed MOIs, bacterial biofilm formation was significantly reduced (P < 0.0001, except for MOI 0.0001 as P = 0.0039) in comparison to the untreated culture after 48 h of incubation ( Figure 6A ). The highest antibiofilm activity was achieved at MOI 100 (OD: 0.32); however, there was no significant difference (P > 0.24) in the results at MOIs ranging from 100 to 0.01. Therefore, MOI 0.01 was considered the lowest MOI with the highest inhibitory effect against biofilm formation.

The capacity of phage vB_Kpn_ZCKp20p to degrade mature biofilm, reflecting its invasion efficiency, was evaluated. A 48-hour-old biofilm was exposed to different phage MOIs, which succeeded in subverting the biofilm after 24 h of all treated cultures, as their ODs were significantly lower (P< 0.0001) than the OD of untreated culture with a mature biofilm ( Figure 6B ). The lowest OD (0.75) was observed at MOI 0.01, which was not significantly different (P > 0.5) when compared to the results at MOIs 1, 0.1, 0.001 and 0.0001. In summary, 0.01 was the lowest MOI of phage vB_Kpn_ZCKp20p with the highest antibiofilm activity to inhibit biofilm formation and disrupt the mature biofilm.

The toxicity of phage vB_Kpn_ZCKp20p was evaluated in vitro by the MTT assay on a common type of human cells (HSF). The MTT assay results did not demonstrate any significant reduction in the cells’ viability after they were exposed to the phage for 24 h and 48 h, which ruled out the cytotoxicity of all assayed phage concentrations (10⁶ PFU/ml to 10⁹ PFU/ml) ( Figure 7 ).