Isolation and characterization of novel bacteriophages with broad-spectrum lytic activity against carbapenem-resistant Klebsiella pneumoniae capsular pathotypes K1, K2, K3, K5, and K20

Abstract

Background

Carbapenem-resistant Klebsiella pneumoniae (CRKP), particularly strains associated with virulent capsular pathotypes K1, K2, K3, K5, and K20, represent a major global health burden due to limited therapeutic options and poor clinical outcomes. This study aimed to isolate and characterize lytic bacteriophages targeting clinically relevant CRKP capsular types.

Methods

Bacteriophages were isolated from hospital sewage and purified by successive plaque assays. Biological characterization included adsorption kinetics, one-step growth analysis, and host range determination. TEM was used for morphological classification. Molecular characterization of the phages was achieved using whole-genome sequencing (WGS).

Results

Three phages, designated UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3, were isolated. The phages exhibited a high and rapid adsorption rate of ~ 95% within 9–10.5 min, short latent periods (~ 25 min), and large burst sizes exceeding 300 PFU/infected cell. All phages efficiently lysed CRKP capsular types K1, K2, K3, K5, and K20. TEM analysis revealed that the phages exhibited siphovirus-like morphology, characterized by similar capsid–tail architecture with polyhedral heads and long non-contractile tails. Based on the current ICTV taxonomy, they were classified within the class Caudoviricetes and the family Drexlerviridae. WGS showed that UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 possess nearly identical genome sizes of 51,620 bp, 51,622 bp, and 51,622 bp, respectively, with high intra-group average nucleotide similarity (> 99% ANI) and were devoid of tRNA genes, lysogenic markers, virulence factors, and antibiotic-resistance determinants. Phylogenomic and VIRIDIC analyses placed the phages within the family Drexlerviridae and genus Webervirus, and revealed that the phages represent novel species within the Webervirus genus.

Conclusion

Therefore, UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 demonstrated strong lytic activity, favorable replication dynamics, and genomic safety, underscoring their potential as candidates for phage therapy against CRKP.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13062-026-00752-0.

Background

Klebsiella pneumoniae (K. pneumoniae) is an encapsulated Gram-negative opportunistic pathogen implicated as one of the major causes of urinary tract infections (UTI), pulmonary infections, meningitis, and sepsis [1]. Encapsulation offers K. pneumoniae virulence, as capsules inactivate the host immune response and shield the pathogen from phagocytosis [2]. Of the 77 capsular (K) pathotypes, K1, K2, K3, K5, K20, K54, and K57 are clinically more significant as they are highly pathogenic and responsible for the biggest percentage of community-acquired K. pneumoniae infections [3].

In K. pneumoniae, cell wall surface polysaccharides such as capsular polysaccharide (CPS) and lipopolysaccharide (LPS) O-antigen are central determinants of pathogenicity and immune evasion. CPS forms a dense extracellular layer that masks underlying surface structures, significantly reducing complement binding and phagocytosis by macrophages, neutrophils, and dendritic cells. This barrier function enables encapsulated strains to evade host innate immunity, survive in the bloodstream, and establish invasive infections [4]. The composition and structure of CPS vary among capsular types, with hypervirulent K1 and K2 serotypes showing particularly strong resistance to phagocytic uptake and intracellular killing, contributing to their enhanced pathogenic potential. LPS, especially the O-antigen component, further contributes to virulence by interfering with complement activation and modulating toll-like receptor signaling, which can influence inflammatory responses during infection [4]. The combined effects of CPS and LPS on cellular interactions with the host immune system underscore why specific capsular pathotypes are more clinically significant and why they remain major contributors to severe CRKP infections.

Due to the excessive use of antibiotics, the emergence of multidrug-resistant (MDR) pathogenic K. pneumoniae, particularly carbapenem-resistant strains with limited treatment choices, has been reported worldwide. For example, Ssekatawa et al. [5] reported that 69.8% of the carbapenem-resistant K. pneumoniae isolated from Uganda Tertiary hospitals were pathogenic. The situation has been complicated by the acquisition of resistance to last-line drugs such as colistin and tigecycline, used in treating K. pneumoniae carbapenem-resistant infections [6]. This further limits the chemotherapeutic options against carbapenem-resistant K. pneumoniae infections. Thus, this calls for the search for new therapeutics, such as phage therapy, to complement antibiotics.

Bacteriophages (phages) are viruses that infect bacteria and use the bacterial machinery to replicate intracellularly, and the progenies are released from the host cell by lysis. The released virions infect other host bacterial cells, and the cycle continues until all the host cells have been lysed; hence, phages are self-limiting [7, 8]. Phages are highly specific; thus, they recognize and infect specific strains within a species. The stringent specificity enables phages to selectively lyse the target pathogens with no substantial effect on the commensal bacteria; hence, this feature makes them suitable for oral administration. In K. pneumoniae, phage specificity is largely governed by cell wall surface-exposed CPS and LPS O-antigens, which function as primary adsorption receptors. Structural variation and similarities in CPS and LPS among clinically important capsular types, including K1, K2, K3, K5, and K20, directly influence phage attachment, host range, and lytic efficiency, thereby shaping phage-host interactions at the bacterial cell surface [9].

Therefore, phages have the potential to selectively target and inhibit distinct pathotypes based on PCR capsular typing and sequencing, as well as degrade the biofilm characteristically formed by K. pneumoniae [10]. This illustrates why phages are the best alternatives in the management of MDR bacterial infections. Indeed, compassionate use of phages to treat MDR bacterial infections in patients where all classes of antibiotics failed has been reported [11, 12]. To combat antimicrobial resistance in developing countries, developing phage therapy research capacity and technology, and creating a local phage library targeting different pathogenic bacteria for potential use in compassionate phage therapy is imminent.

This study reports three novel, safe, strictly lytic phages isolated from the Mulago National Referral Hospital effluents with potent lytic activity against carbapenem-resistant K. pneumoniae capsular pathotypes K1, K2, K3, K5, and K20 isolated from Ugandan tertiary hospitals [5]. Furthermore, the phages’ genome structure, morphology, size, pH range, temperature range, UV light impact, replication kinetics, and host range are in here.

Materials and methods

Source of bacterial strains

The study used carbapenem-resistant K. pneumoniae capsular pathotypes K1, K2, K3, K5, and K20 initially characterized by [5]. All the isolates were sub-cultured in Brain Heart Infusion Broth (BHIB) and stored at -20 °C until further use. K. pneumoniae strain K10, used as the host for the isolation of phages and other Enterobacteriaceae family members, namely enteropathogenic Escherichia coli (EPEC), Shiga toxin–producing E. coli (STEC), enterotoxigenic E. coli (ETEC), enteroinvasive E. coli (EIEC), uropathogenic E. coli (UPEC), Salmonella enterica, Shigella dysenteriae, and Enterobacter cloacae used in the host range experiments, were obtained from the Microbiology Laboratory, College of Veterinary Medicine, Animal Resources and Biosecurity, Makerere University. Reference strains, including K. pneumoniae ATCC^® BAA-1705, K. pneumoniae strain Kp36, E. coli ATCC 25,922, and E. coli O157:H7, were provided by the Microbiology Laboratory, College of Health Sciences, Makerere University, as a kind gift.

Phage isolation and proliferation

Isolation of phages with lytic efficacy against K. pneumoniae was performed using the double agar overlay assay and spot assay as previously described by Wang et al. [13] and Marashi et al. [14], with some modifications. Briefly, 50 mL of each effluent sample collected from Mulago National Referral Hospital, Uganda, was centrifuged at 5000 rpm for 15 min. The obtained supernatant was syringe filtered through 0.22 μm pore size microfilters (Whatman^® Anotop^®) to remove bacterial cells and sewage debris. This was followed by mixing an equal volume (20 mL) of bacteria-free filtrate and double-strength Brain Heart Infusion Broth (BHIB) supplemented with 1 M calcium chloride. The mixture was enriched with 200 µL log-phase broth culture of K. pneumoniae strain K10 (OD₆₀₀ 0.3 (Approximately 10⁸ CFU/mL)) and then later incubated at 37 °C for 24 h while shaking at 180 rpm. The mixture was centrifuged at 10,000 x g for 10 min to obtain a supernatant, which was syringe filtered through 0.22 μm pore size microfilters. The obtained bacteria-free filtrate (crude phage lysate) was spot assayed to confirm the presence of lytic phages against the host bacteria. Briefly, the overlay Brain Heart Infusion Agar (BHIA) plates were initially prepared by mixing and vortexing 4.4 mL of molten agar (0.7%) previously maintained at 55 °C with 100 µL of an overnight bacterial host culture. The overlay was allowed to dry aseptically at room temperature, spotted with 10 µL of the 10-fold serially diluted filtrate, and then incubated overnight at 37 °C.

Following phage isolation, discrete plaques formed on the bacterial lawn were selected based on plaque morphologies and plugged out of the soft agar using sterile bacteriological loops. Each plaque was resuspended in 20 mL of BHIB containing overnight K. pneumoniae cultures, followed by incubation at 37 °C for 24 h. Chloroform was added to the phage lysate mixture to inactivate the bacteria, and then the overlay-spot assay was performed as described above to purify each phage. Three rounds of purification were performed to obtain pure stocks of each isolated phage.

Proliferation of each pure monoclonal phage was done using a slightly adjusted double overlay agar method previously described by Jo et al. [15]. Briefly, the double overlay plaque assay was performed by thoroughly mixing 100 µL of each pure phage with 4.4 mL of molten agar (0.7% agar) containing the host bacterial cells at OD₆₀₀ 0.3 (Approx 10⁸ CFU/mL). This mixture was poured onto the BHIA plate, left to solidify at room temperature, and then incubated at 37 °C for 24 h. Using a sterile spatula, the lysate layer of the overlay was scraped off, resuspended in 20 mL sodium magnesium (SM) buffer (2% gelatin, 5 mL; 1 M Tris-Cl pH 7.5, 50 mL; MgSO₄.7H₂O, 2 g; NaCl, 5.8 g, and 100mL of distilled deionized water). The phage lysate suspended in SM buffer was syringe filtered through a 0.22 μm pore size microfilter to remove the bacterial debris, and the obtained filtrate was stored at 4° C until further use.

Estimation of the plaque-forming units

The concentration of phages in a solution was estimated by spot assay of ten-fold serial dilution to enumerate the plaque-forming units per millilitre (PFU/mL) as described by Mugo et al. [16]. Host bacterial lawns were made using an overlay assay prepared by mixing 100 µL of an overnight K. pneumoniae culture with 4.4 mL of molten 0.7% agar, which was then poured onto BHIA plates and allowed to solidify. The spot assay was performed by spotting in triplicate 10 µL of serially diluted (10-fold) phages in SM buffer till they air-dried at room temperature. The spotted plates were incubated overnight at 37 °C, and all visibly clear plaques of each phage were counted for each dilution factor. Experiments for each phage were performed in triplicate. The dilution factors that contained more than ten plaques were considered for enumeration. The PFU/mL was calculated using formula 1.

1

Phage purification

Phage purification and concentration were done using the glycerol gradient ultracentrifugation method previously described by Yazdi et al. [17]. Briefly, 10 mL of each phage in SM buffer (approximately 10¹⁰ PFU/mL) was centrifuged at 10,000 × g for 20 min, and then the obtained supernatant was syringe-filtered through a 0.22 μm pore size microfilter. The filtrate was treated with RNase (1 µg/µL) and DNase (1 µg/µL), then incubated at 37 °C for 1 h to eliminate any bacterial genetic material. Subsequently, 10% polyethylene glycol (PEG) 8000 and Sodium chloride (1 M) were added to the phage suspension and stored overnight at 4 °C. The phages were precipitated from the suspension by centrifugation at 13,000 × g for 30 min at 4 °C to obtain a pellet, which was later resuspended in 2mL of SM. Further purification of the concentrated phage suspension was done using a glycerol step gradient of 40% SM buffer and 5% glycerol, followed by ultracentrifugation at 15,000 × g for 2 h and at 4 °C (Beckman ultracentrifuge, SW28 rotor). The formed phage precipitates were resuspended in SM buffer and then kept at 4 °C until further use.

Ultrastructural characterization by transmission electron microscopy

The morphology of the three phages was examined using transmission electron microscopy (TEM) at the Advanced Bioimaging Facility, University of Warwick, under the technical support of Dr. Saskia Bakker. For sample preparation, 5 µL of 10⁹ PFU/ml phage lysate was deposited onto glow-discharged (1 min under vacuum) formvar–carbon-coated copper grids (Agar Scientific Ltd., UK) and allowed to adsorb at room temperature for approximately 90 s. Excess fluid was carefully removed by a bloating paper, after which the grids were negatively stained using 2% (w/v) uranyl acetate for 1 min. This staining step was repeated four times to enhance contrast, followed by air-drying. The prepared grids were examined using a JEOL 2100Plus transmission electron microscope. Digital micrographs were processed and analyzed using ImageJ software, and virion dimensions were calculated based on measurements from at least ten individual phage particles, with mean values reported.

Determination of the host range

The host range of the isolated phages was determined using the spot assay as previously described by Fikadu et al. [18] with slight modifications. Briefly, lawns of each test bacteria type were prepared on BHIA plates using the double agar overlay method. Each plate was divided into equal zones and clearly labelled with the corresponding phage isolate and dilution factor. This was followed by 10-fold serial dilution of each phage in SM buffer, and subsequently spotted 10 µL of each dilution on the bacterial lawn, allowed to air-dry at room temperature, and then incubated overnight at 37 °C. This step was repeated thrice. K. pneumoniae pathotype plates with clear and discrete plaques were considered to be susceptible to phage-mediated lysis. Still, based on the level of clarity on the bacterial lawn, the spots alongside the formed plaques were classified as highly clear plaques measuring 2.5–3.0 mm (+++), clear plaques ranging from 2.0 to 2.49 mm (++), partially turbid plaques with diameters of 1.2–1.99 mm (+), or absence of detectable lysis (-). Halo diameter, a translucent zone surrounding the central lytic plaque, was measured. The total diameter of the halo plus plaque was measured, and the plaque diameter was subtracted to determine the halo width. Measurements were performed after incubation and averaged from three independent replicates.

Efficiency of plating (EOP)

The efficiency of plating was evaluated for phages that exhibited a broader host range to further assess the productive infectivity using the double agar overlayer plaque assay described by Mirzaei & Nilsson [19]. Briefly, 100 µL of diluted phage lysate (10⁶–10⁹) of each phage was mixed with 200 µL of test bacterial strains and host bacterium at OD₆₀₀ 0.3 (Approx. 10⁸ CFU/mL) contained in separate Eppendorf tubes and then incubated for a sufficient adsorption time at 37 °C. The phage-bacteria mixture was transferred into 4.4 mL of molten agar (0.7%), mixed by vortexing prior to plating on BHIA plates. Later, the plates were incubated overnight at 37 °C. The assays were performed in triplicate to calculate the average phage titres for each bacterium, as shown by formula 2.

2

The EOP for each phage-bacterium combination was interpreted as follows: high production efficiency (≥ 0.5 or 50%), medium production efficiency (≥ 0.1 (10%) to < 0.5 (50%)), low production efficiency (0.001 (0.1%) to > 0.1(10%)) and inefficient production (≤ 0.001 (0.1%)) [16].

Effect of temperature, pH, and UV light on phage stability

The stability of phages at various temperatures, pH, and UV Light was examined using the double agar overlay plaque method previously described by Petrusic et al. [20] and Ssekatawa et al. [21]. For thermal stability, phage lysates were incubated at − 20, 4, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, and 80 °C in a temperature-controlled water bath for 1 h. Briefly, 1 mL of Tryptic Soy Broth (TSB) with a titre of 10⁸ PFU/mL for each phage was placed into corresponding Eppendorf tubes and incubated for 1 h at their corresponding temperatures. Aliquots of 100 µL were pipetted at 10-minute intervals in triplicate and mixed with 100 µL of the host bacterial suspension (10⁸ CFU/mL). The mixture was aseptically added to 4.4 mL of molten agar (0.7% Tryptic Soy Agar (TSA)), vortex-mixed, and then overlaid onto prepared TSA plates for phage titre enumeration. The phage viability and concentration were evaluated using the soft agar overlay followed by the plaque assay, and titres were expressed as PFU/mL. The same procedure was utilized for UV light tests; however, for UV stability tests, incubation was done at 37 °C, and aliquots of 100 µL were removed at 0, 5, 10, 20, 30, 40, 50, and 60 min in triplicate. For pH stability assessment, 1 mL of each phage lysate (10⁸ PFU/mL) was exposed to pH values ranging from 2 to 14 and incubated for 1 h at 37° C. Following incubation, residual phage viability was determined in triplicate using the double-layer agar (soft agar overlay) plaque assay, as described under the temperature stability test. The phage viability and survival rate of each phage in the described stability studies were expressed as the ratio of the obtained final phage titer to the initial phage titre.

Phage adsorption period, latent period, and burst size

The phage adsorption assay of three phages was conducted following the methods previously described by Ahmad et al. [22], but with some slight modifications. Briefly, 2 mL of an overnight culture of K. pneumoniae K10 13883 in tryptic soy broth (TSB) was inoculated into 18 mL of TSB, incubated at 37 °C until a density of OD₆₀₀ 0.3 (Approx 10⁸ CFU/mL) was achieved. Approximately 100µL of each phage stock (10⁹ PFU/mL) was mixed separately in sterile 15 mL Falcon tubes with the host bacteria at a multiplicity of infection (MOI) of 0.1 and incubated at 37 °C for 45 min. At a 2.5-minute interval, 100 µL aliquots were collected in duplicates throughout the duration of the experiment. Each sample was diluted in 4.4 mL TSB, followed by treatment with 0.5 mL of chloroform to inactivate any bacteria. Subsequently, the samples were centrifuged at 12,000 X g to sediment the bacterial debris, harvested the supernatant containing unadsorbed phages was later passed through 0.22 μm filters. The titres of the non-adsorbed phages at each time interval were enumerated via double-layer agar assay and were used to calculate the percentage of non-adsorbed phages as the ratio of phage titres in the harvested supernatant to the initial titres. The percentage of the non-adsorbed phages was plotted against the time intervals to generate the adsorption curve.

A one-step growth assay was performed to determine the latent period and burst size of phage using the methods previously outlined by Lin et al. [23] but with minor modifications. An initial bacteria-phage suspension at a MOI of 0.1 was incubated at 37 °C to allow sufficient adsorption of phages to the bacterial cells for a duration based on the estimated based on the estimated adsorption time previously described. Subsequently, the suspension was centrifuged at 12,000 × g for 5 min, and the resulting pellet containing the infected bacterial cells was immediately resuspended in 10 mL of prewarmed (37 °C) TSB and later incubated at 37 °C for 120 min. At 5 min intervals, 100 µL of each sample was withdrawn in duplicates, followed by a 10-fold serial dilution and titration using a double-layer agar assay. The calculated titres obtained at different time intervals were used to plot a one-step growth curve. From this curve, the latent period was determined as the interval between phage adsorption and the subsequent release of progeny phage particles. The burst size was calculated as the ratio of the final free phage titre to the initial phage titre.

Genomic DNA extraction, sequencing, and bioinformatic analysis

Phage DNA was isolated using the Phage DNA Isolation Kit (Norgen) from 10⁹ PFU/ml of purified phage lysates following the manufacturer’s instructions. Construction of short read paired-end sequencing libraries and whole-genome sequencing were performed by MicrobesNG (Birmingham, UK) on an Illumina platform, generating 2 × 250 bp reads. Genome assembly and annotation were conducted in accordance with recommended phage genome annotation workflows [24]. Raw reads were quality-filtered and trimmed using Trimmomatic [25] before de novo assembly with SPAdes genome assembler v2.5.1 [26] using the --only-assembler option. The resulting contigs were annotated using Prokka v1.11 [27], which incorporated the PHROG database [28] and a custom phage genome database compiled from GenBank entries [29]. Furthermore, FASTA files containing phage protein sequences were scanned for putative depolymerases using DePP, a DePolymerase Predictor tool. Simultaneously, phage protein sequences were compared against a list of known depolymerase sequences using Blastp. The annotated genome sequences were deposited in the European Nucleotide Archive (ENA) under accession numbers: UGKSKpnP1 (OV877387), UGKSKpnP2 (OV877339), and UGKSKpnP3 (OV877766).

Comparative phylogenomics and evolutionary analysis

The reconstructed phage genomes were first examined using ViPTree to infer their global relatedness to reference phages [30]. To identify the closest known relatives, whole-genome BLASTn comparison searches were conducted against the NCBI phage database. Genomes of phages assigned to the same viral subfamily were retrieved and uniformly re-annotated using Prokka v1.11 [27] with the previously described custom database. Conserved marker genes, specifically the major capsid protein or the large terminase subunit, were identified using USEARCH v10.0.240 with an e-value threshold of 1e-5 [31]. Corresponding nucleotide sequences were aligned with mafft v7.271 [32], and phylogenetic trees were inferred using IQ-TREE v1.6.3 with 1,000 ultrafast bootstrap replicates, applying evolutionary models selected by MODELTEST [33, 34]. For genus-level resolution, core genome analysis was performed using ROARY with --e --mafft -p 32 –i 90 parameters [35]. Phylogenies based on concatenated core genes were reconstructed using IQ-TREE as described above. Pairwise genome similarity and average nucleotide identity were calculated with VIRIDIC to assess taxonomic novelty at the genus level [36]. The inferred phylogenetic trees were graphically represented using iTOL, Interactive Tree of Life [37].

Data analysis

Data were analyzed using OriginPro 2021. Differences in phage replication dynamics, including adsorption efficiency, latent period, and burst size, as well as phage stability under varying pH, temperature, and UV light exposure, were assessed using one-way ANOVA followed by Tukey’s multiple comparison tests. Values of P ≤ 0.05 were considered statistically significant.

Ethics statement

Ethical approval No. MHREC1611) was obtained from the Research and Ethics Committee at Mulago National Referral Hospital. The Committee waived the requirement for informed consent, as this study utilized pre-coded archived samples. All experimental procedures were conducted in accordance with the approved ethical guidelines.

Results

Isolation, purification, and amplification of bacteriophages

Three lytic phages infecting K. pneumoniae were isolated from hospital effluents using K. pneumoniae strain K10 as the host bacteria. The phages were designated Uganda Kenneth Ssekatawa K. pneumoniae phages 1–3 (UGKSKpnP1-UGKSKpnP3). Among them, UGKSKpnP1 produced the largest plaques, with a mean diameter of 3 mm; however, plaque sizes did not differ significantly among the three phages (p = 0.999). All isolates formed pronounced halos extending up to 16 mm in diameter after 24 h of incubation (Fig. 1A1, B1, and C1, Table 1). Following purification and amplification, the phages attained high titers, each reaching approximately 6.4 × 10¹² PFU/mL.

Fig. 1.

Representative images of Klebsiella pneumoniae phage isolates: A1, B1, and C1; spot assays demonstrating lytic activity of UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3, respectively, and A2, B2, and C2; TEM micrographs illustrating the morphology of UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3, respectively

Table 1.

Lytic spectrum of the bacteriophages against clinically and epidemiologically important Enterobacteriaceae

Host bacteria	phage host range						EOP (%)
	UGKSKpP1		UGKSKpP2		UGKSKpP3		UGKSKpP1	UGKSKpP2	UGKSKpP3
	Plaque	Halo (mm)	Plaque	Halo (mm)	Plaque	Halo (mm)
K. pneumoniae capsular pathotypes K1	+++	15	+++	14	+++	14	100	98	95
K. pneumoniae capsular pathotypes K2	+++	16	+++	16	+++	15	100	100	100
K. pneumoniae capsular pathotypes K3	+++	15	+++	15	+++	15	100	94	100
K. pneumoniae capsular pathotypes K5	+++	15	+++	15	+++	15	100	100	90
K. pneumoniae capsular pathotypes K20	+++	15	+++	13	+++	15	100	90	99
K. pneumoniae capsular pathotypes K10	+++	16	+++	16	+++	15	100	100	100
K. pneumoniae ATCC® BAA-1705	+++	15	+++	16	+++	15	100	100	100
K. pneumoniae strain Kp36	++	13	++	10	++	11	70	65	59
Escherichia coli ATCC 25,922	-	0	-	0	-	0	0	0	0
E. coli O157:H7	-	0	-	0	-	0	0	0	0
Enteropathogenic E. coli (EPEC)	+	7	+	5	+	8	35	35	35
Shiga toxin-producing E. coli (STEC)	+	7	+	7	+	7	50	50	50
Enterotoxigenic E. coli (ETEC)	+	7	+	7	+	7	28	28	28
Enteroinvasive E. coli (EIEC)	+	5	+	7	+	6	40	40	40
Uropathogenic E. coli (UPEC)	+++	12	+++	10	+++	14	95	95	95
Salmonella enterica	-	0	-	0	-	0	0	0	0
Shigella dysenteriae	-	0	-	0	-	0	0	0	0
Enterobacter cloacae	-	0	-	0	-	0	0	0	0

In this table, lytic activity was assessed based on plaque or spot appearance and diameter on the bacterial lawn. Reactions were categorized as highly clear plaques measuring 2.5–3.0 mm (+++), clear plaques ranging from 2.0–2.49 mm (++), partially turbid plaques with diameters of 1.2–1.99 mm (+), or absence of detectable lysis (-)

TEM-based virion morphology and classification

Phage classification was performed in accordance with the guidelines of the International Committee on Taxonomy of Viruses (ICTV), using ultrastructural features observed under transmission electron microscopy. Examination of UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 revealed highly comparable virion morphologies. The virions possessed well-defined polyhedral capsids with average dimensions of approximately 62 ± 7 nm in length and 56 ± 4 nm in width, attached to elongated, flexible, non-contractile tails measuring about 169 ± 9 nm in length and 10 ± 1 nm in width (Fig. 1A2, B2, C2).

Host range specificity assessment

The host range of the three K. pneumoniae phages isolated on K. pneumoniae strain K10 was examined using spot assays. The phages were tested against a panel of carbapenem-resistant K. pneumoniae isolates representing capsular types K1, K2, K3, K5, and K20 that had been recovered from clinical samples. In addition, other clinically and epidemiologically relevant members of the family Enterobacteriaceae were included, namely enteropathogenic Escherichia coli (EPEC), Shiga toxin–producing E. coli (STEC), enterotoxigenic E. coli (ETEC), enteroinvasive E. coli (EIEC), uropathogenic E. coli (UPEC), Salmonella enterica, Shigella dysenteriae, and Enterobacter cloacae. Reference strains, including K. pneumoniae ATCC^® BAA-1705, K. pneumoniae strain Kp36, E. coli ATCC 25,922, and E. coli O157:H7, were also incorporated into the analysis. The assay results demonstrated that all three phages were able to lyse each of the tested K. pneumoniae capsular types, as well as K. pneumoniae ATCC^® BAA-1705 and UPEC isolates, producing highly distinct and clear plaques. Clear plaque formation was also observed against the hypervirulent K. pneumoniae strain Kp36, whereas only partially turbid lysis zones were detected on EPEC, STEC, ETEC, and EIEC lawns. In contrast, no lytic activity was detected against Salmonella enterica, Shigella dysenteriae, Enterobacter cloacae, E. coli ATCC 25,922, and E. coli O157:H7 (Table 1).

Efficiency of plating (EOP) analysis

The efficiency of plating analysis showed that the three phages (UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3) exhibited different levels of infectivity across the susceptible bacterial strains tested. The EOP values ranged from 28% to 100% (Table 1). Tukey’s multiple comparison test showed that EOP values were significantly higher for all K. pneumoniae capsular types (K1, K2, K3, K5, and K20), K. pneumoniae ATCC^® BAA-1705, and UPEC isolates than for the hypervirulent K. pneumoniae strain Kp36 (p = 0.045) and the Escherichia coli pathotypes EPEC, STEC, ETEC, and EIEC (p = 0.0025).

Temperature, pH, and UV light stability of phages

The thermal tolerance of the three phages was evaluated by incubating phage suspensions at different temperatures. Phage viability remained unchanged across the temperature range from − 20 °C to 35 °C, with titers maintained at 100%. A marked decline in phage viability was observed as the temperature increased from 40 °C to 65 °C, where survival decreased progressively from approximately 95% to 0.001%, representing a statistically significant difference (p = 0.0002). Complete loss of infectivity occurred at 70 °C (Fig. 2).

Fig. 2.

Thermal stability of UGKSKpn phages. Viability and percentage survival of UGKSKpnP1 (A, B), UGKSKpnP2 (C, D), and UGKSKpnP3 (E, F) following exposure to different temperature conditions

The influence of pH on phage stability was examined by exposing the phages to pH conditions ranging from 2 to 14 at 37 °C for 60 min. Thereafter, each treated sample was subjected to the double-layer agar (soft agar overlay) plaque assay to enumerate residual phage viability following pH exposure. Phage activity increased with rising pH and peaked near neutrality, with mean survival of 23.3 ± 11.5, 23.6 ± 11.5, and 29.0 ± 11.9 for UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3, respectively. Minimal activity was detected in phages that were exposed to pH 2, which was significantly lower than the maximal activity observed in phages treated with pH 7.45 (p = 0.001), where phage titers exceeded the initial reference value of 100% (10⁹ PFU/ml) (Fig. 3A–F). Beyond pH 9, phage viability declined sharply (p = 0.0111), reaching a minimum of approximately 10³ PFU/mL, corresponding to a survival rate of 0.001%. No significant differences in pH tolerance were observed among the three phages (P = 0.92237, and F = 0.8096) (Fig. 3G and H).

Fig. 3.

pH stability of UGKSKpn phages. Viability and percentage survival of UGKSKpnP1 (A, B), UGKSKpnP2 (C, D), and UGKSKpnP3 (E, F) following exposure to different pH conditions. Panels (G, H) summarize descriptive statistics of pH-dependent survival among UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3; Panel G represents the mean phage survival at different pH treatments, and Panel H shows the mean phage survival at pH 7.37. Error bars represent variability among replicates (Same-sized error bars and letters indicate similar variability among groups), and the absence of statistically significant differences is indicated where p > 0.05

To assess UV sensitivity, the phages were exposed to ultraviolet light at 240 nm. Phage infectivity decreased rapidly with increasing exposure time, with titers declining from 10⁹ PFU/mL to approximately 10⁴ PFU/mL within 10 min. Complete inactivation was observed after 40 min of exposure, resulting in a survival rate of 0% (Fig. 4A–C). The response to UV light exposure did not differ significantly among UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 (p = 0.99726) (Fig. 4D).

Fig. 4.

UV light stability of UGKSKpn phages. Percentage survival of UGKSKpnP1 (A), UGKSKpnP2 (B), and UGKSKpnP3 (C) following exposure to UV light. Panel (D) summarizes descriptive statistics of UV-dependent survival among UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3. Error bars represent variability among independent replicates (mean ± SD)

Phage replication dynamics

The replication characteristics of UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 were assessed through one-step growth curve analysis, focusing on adsorption efficiency, eclipse/latent phase duration, and progeny yield (Fig. 5). For UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3, the minimum titre value of unbound phage particles was detected at approximately 9, 10, and 10.5 min, respectively. Despite these slight temporal differences, adsorption efficiencies were comparable among the three phages, ranging between 94.5% and 95.1%, with no statistically significant variation (p = 0.9999). Furthermore, all three phages exhibited a latent phase of approximately 25 min before the onset of intracellular phage release. The average number of progeny virions produced per infected bacterial cell was also similar, with estimated burst sizes of about 321, 311, and 340 PFU per host bacterial cell for UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3, respectively. Statistical analysis confirmed that differences in burst size among the phages were not significant (p = 0.9899), indicating comparable replication efficiencies across the three isolates.

Fig. 5.

Phage adsorption kinetics and one-step growth curves of UGKSKpn phages. Panels (A–C) show the adsorption rates of UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3, respectively. Panels (D–F) depict the corresponding one-step growth curves for UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3, indicating their latent periods and burst sizes

Genome analysis and annotation

Genome assembly and features

Phages UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 were assembled into single contigs of 51,620 bp, 51,622 bp, and 51,622 bp, respectively (Fig. 6), with an average nucleotide identity (ANI) of 99.43% among them. The genomes encoded 79 predicted coding sequences (CDS) for UGKSKpnP1 and 80 for UGKSKpnP2 and UGKSKpnP3. Of these, 37 CDS in UGKSKpnP1 and UGKSKpnP2, and 38 in UGKSKpnP3 were assigned putative functions, while 42 CDS in UGKSKpnP1 and UGKSKpnP3, and 43 in UGKSKpnP2 were annotated as hypothetical proteins of unknown function (Supplementary Table 1). No tRNA genes were identified. Functional annotation confirmed the absence of bacterial virulence factors, toxins, toxin-related genes, and antibiotic resistance determinants. Furthermore, integrase genes or other markers of a temperate/lysogenic lifestyle were not detected.

Fig. 6.

Comparative genomic organization of Klebsiella pneumoniae phages. Schematic genome maps illustrating genome sizes and the distribution of selected predicted genes for UGKSKpnP1 (A), UGKSKpnP2 (B), and UGKSKpnP3 (C)

Phage-encoded depolymerases

DePP scan analysis identified several phage proteins as potential depolymerases, with a subset of annotated “tail proteins” and “tail fiber proteins” ranking among the top 10 predicted hits (Table 2). BLASTp analysis further demonstrated that all three analyzed phages encoded proteins highly similar to previously characterized depolymerases. The putative depolymerases of UGKSKpnP1 (OV877387) and UGKSKpnP3 (OV877766) showed high similarity to the depolymerase of Klebsiella phage KP36 targeting capsule type K63 (97% query coverage, 90% amino acid identity) and to a putative depolymerase of Klebsiella phage KLPN1 targeting capsule type K1 (100% query coverage, 94% amino acid identity). For UGKSKpnP2 (OV877339), initial analysis identified only a protein similar to the Klebsiella phage KLPN1 capsule type K1 depolymerase. However, further genome inspection revealed that protein OV877339_00080 exhibited similarity to the KP36 K63 depolymerase (85% query coverage, 81% identity). Following genome circularization and reordering of the gene arrangement, the complete open reading frame was restored (Supplementary data 1), confirming that UGKSKpnP2 encodes a full-length depolymerase homologous to that of phage KP36.

Table 2.

Top 10 DePP hits and their respective Blastp search results for each phage’s proteins. Green colour denotes putative functional depolymerases, yellow colour denotes truncated protein, and dark green colour denotes reinstated protein

UGKSKpnP2-OV877339				Best Blastp hit results
Protein locus tag	Annotation	Length (amino acid)	Probability Depp	Target (type_name_accession _capsule target)	Coverage	Evalue	Identity	Length (amino acid)
OV877339_00078	tail fiber protein	877	0.963270635	depo_KlebsPhageB1_QTP95996_K2	26%	3.00E-70	61.37%	907
OV877339_00005	tail length tape measure protein	988	0.862607864	n/a
OV877339_00072	DNA helicase	678	0.746656277	n/a
OV877339_00079	membrane protein	318	0.639971645	n/a
OV877339_00019	terminase large subunit	534	0.627213685	n/a
OV877339_00075	exonuclease VIII	348	0.579812168	n/a
OV877339_00058	DNA helicase	522	0.576001056	n/a
OV877339_00018	portal protein	439	0.574830951	n/a
OV877339_00008	major tail protein	228	0.415386386	depo_KLPN1_YP09195374_K1	100%	1.00E-167	93.86%	228
OV877339_00002	minor tail protein	245	0.391160845	n/a
OV877339_00080	hypothetical protein	204	n/a	depo_KP36_YP009226010_K63	85%	5.00E-93	81.03%	1233
OV877339_00080reinstated	tail protein	1150	n/a	depo_KP36_YP009226010_K63	97%	0.00E + 00	90.09%	1233
UGKSKpnP1-OV877387
OV877387_00018	tail protein	1150	0.979422222	depo_KP36_YP009226010_K63	97%	0	90.09%	1233
OV877387_00020	tail fiber protein	877	0.966009524	depo_KlebsPhageB1_QTP95996_K2	26%	3.00E-70	61.37%	907
OV877387_00013	tail length tape measure protein	988	0.866464550	n/a
OV877387_00026	DNA helicase	678	0.763306878	n/a
OV877387_00019	membrane protein	318	0.648491534	n/a
OV877387_00079	terminase large subunit	534	0.611731217	n/a
OV877387_00040	DNA helicase	522	0.581264815	n/a
OV877387_00023	exonuclease VIII	348	0.556381217	n/a
OV877387_00010	major tail protein	228	0.412667196	depo_KLPN1_YP09195374_K1	100%	7.00E-167	93.42%	228
OV877387_00002	head maturation protease	376	0.377110847	n/a
UGKSKpnP3-OV877766
OV877766_00013	tail protein	1050	0.981561111	depo_KP36_YP009226010_K63	97%	0	90.09%	1233
OV877766_00011	tail fiber protein	877	0.958584392	depo_KlebsPhageB1_QTP95996_K2	26%	3.00E-70	61.37%	907
OV877766_00018	tail length tape measure protein	988	0.868532011	n/a
OV877766_00005	DNA helicase	678	0.781475397	n/a
OV877766_00012	membrane protein	318	0.659311376	n/a
OV877766_00032	terminase large subunit	534	0.631870635	n/a
OV877766_00031	portal protein	439	0.579063059	n/a
OV877766_00071	DNA helicase	522	0.577791005	n/a
OV877766_00008	exonuclease VIII	348	0.562225397	n/a
OV877766_00021	major tail protein	228	0.419036243	depo_KLPN1_YP09195374_K1	100%	1.00E-167	93.86%	228

Phylogenetic classification

Based on total protein content, ViPTree analysis revealed that UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 clustered with the Caudoviricetes class and the Drexlerviridae family phages that infect Gammaproteobacteria (Fig. 7). Blastn comparison further showed high similarity (> 96% identity) to Klebsiella phages: vB_KpnS_Domnhall, Klebsiella phage vB_KpnS_ZX2, Klebsiella phage 066039, and other members of the Drexlerviridae family. A phylogenetic tree based on the terminase large subunit sequences of UGKSKpnP1, UGKSKpnP2, UGKSKpnP3, and Drexlerviridae family phages confirmed their placement within the genus Webervirus (Fig. 8).

Fig. 7.

Proteome-based phylogenetic analysis of bacteriophages. The phylogenetic tree was constructed using ViPTree and genome-wide protein sequence similarities. Phages UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 are highlighted with star symbols. The analysis includes all available reference phage genomes

Fig. 8.

Major capsid protein–based phylogeny of Drexlerviridae phages. The phylogenetic tree was inferred from major capsid protein sequences using the LG + R4 substitution model with 1,000 bootstrap replicates. Phages UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 are highlighted by a green background. Filled circles indicate bootstrap support values greater than 70%, with circle size proportional to the level of bootstrap support. Colored outer rings denote phage genera

Core gene phylogeny of Webervirus phages

To determine their taxonomic position, a phylogenetic analysis using core genes from UGKSKpnP1, UGKSKpnP2, UGKSKpnP3, and all Webervirus phages was conducted. The analysis showed that the three phages formed a distinct clade within the genus, closely related to Klebsiella phage vB_KpnS_Call and Klebsiella phage KOX1, which are already classified as Klebsiella virus Call and Klebsiella virus KOX1 species, respectively, by ICTV (Fig. 9). These findings suggest that UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 likely represent novel and yet uncharacterized phage species.

Fig. 9.

Core gene-based phylogeny of Webervirus phages. The phylogenetic tree was reconstructed using concatenated nucleotide sequences of 13 core genes under the GTR + F+R4 evolutionary model with 1,000 bootstrap replicates. Phages UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 are highlighted with a green background. Node support values are indicated by circles at branch points: small filled circles represent bootstrap support ≥ 70%, whereas larger circles indicate bootstrap values approaching 100%, with circle size proportional to the magnitude of the bootstrap value. Branch lengths correspond to the number of nucleotide substitutions per site

Species delineation using VIRIDIC

VIRIDIC analysis further supported their taxonomic placement, showing that UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 shared a maximum of 93.2% intergenomic similarity with other analyzed Webervirus phages, which is below the 95% species demarcation threshold. The three genomes clustered together within a distinct species cluster (Cluster 45), separate from all previously described species (Supplementary Tables 2 and 3). Based on the accepted 95% ANI cutoff for species delineation, UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 represent a single novel species within the genus Webervirus (Supplementary Table 3, Fig. 10).

Fig. 10.

VIRIDIC-based intergenomic similarity analysis of Webervirus phages. The heatmap displays pairwise intergenomic similarity values, with darker shading indicating higher nucleotide similarity. The right half of the matrix presents similarity percentages, while the left half shows alignment metrics, including the fraction of the genome aligned for the row genome (top value), the genome length ratio between genome pairs (middle value), and the fraction of the genome aligned for the column genome (bottom value). Genome sizes are represented by grey bars positioned above the heatmap

Discussion

Klebsiella pneumoniae infections are largely driven by multidrug-resistant and hypervirulent strains, many of which are differentiated by distinct CPS types that play critical roles in immune evasion and pathogenicity. Therefore, assessment of phage host range across prevalent capsular types is a key requirement when selecting candidate phages for phage therapy and other biocontrol applications. In this study, three lytic phages (UGKSKpnP1-UGKSKpnP3) were isolated and evaluated for their lytic spectrum against clinically relevant K. pneumoniae strains and related Enterobacteriaceae. All three phages demonstrated strong lytic activity against carbapenem-resistant K. pneumoniae isolates representing capsular types K1, K2, K3, K5, and K20, as well as the reference strain ATCC^® BAA-1705 and the hypervirulent clinical strain Kp36. The formation of large, clear plaques with high EOP across these capsular variants indicates efficient adsorption, replication, and host cell lysis, suggesting that these phages either recognize conserved surface receptors or possess enzymatic mechanisms that enable penetration of structurally diverse capsules, hence polyvalent. This phenomenon has been of rare occurrence; however, several recent studies have reported polyvalent phages that exhibit remarkable host diversity. For example, Dhungana et al. [38] isolated E. coli phages that could also lyse different K. pneumoniae strains. Similarly, several studies have documented phage intergeneric and inter-family lytic activities [39–41]. The prominent halo formation surrounding the plaques is due to the presence of phage-associated polysaccharide-degrading enzymes, such as depolymerases, which are often implicated in capsular polysaccharide degradation and may contribute to enhanced host accessibility and broad-spectrum lytic activity [9, 42, 43]. Indeed, this study revealed that all three phages contain multiple putative depolymerases, which are similar to known depolymerases targeting K1 and K63 Klebsiella capsule types. Further functional analysis is required to confirm the protein activity and their capsular targets. Regarding phage therapy, phages capable of infecting hosts across multiple bacterial genera functionally resemble broad-spectrum antimicrobials [44]. Such polyvalent phages are highly advantageous, as they reduce the need for isolating and formulating strain-specific phages for individual infections, thereby simplifying therapeutic design and expanding phage-mediated applicability. However, only partial or turbid lysis with low EOP was observed on lawns of DEC pathotypes, including EPEC, STEC, ETEC, and EIEC. Such incomplete lysis may reflect non-productive phage-host interactions, potentially arising from bacteriocin-mediated antagonism or disruption of host cellular energetics that results in abortive infection rather than successful phage replication [45]. In total contrast, no activity was detected against Salmonella enterica, Shigella dysenteriae, Enterobacter cloacae, or standard E. coli reference strains. This restricted cross-genus activity highlights the dominant role of specific surface receptors, such as CPS, outer membrane proteins, and lipopolysaccharides, in governing phage specificity [46]. The susceptibility of UPEC isolates to the three K. pneumoniae phages, but not reference E. coli strains, may reflect shared functionally analogous surface adhesion structures between UPEC and K. pneumoniae. Such similarities, potentially arising from conserved features such as fimbrial and polysaccharide components in Gram-negative bacteria, are known to influence phage receptor recognition and may therefore contribute to host range overlap among related taxa [46].

TEM analysis of the representatives of UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 revealed that the phages were tailed, thus belonging to the class Caudoviricetes [47, 48]. Furthermore, the phages had visible polyhedral heads (length: 62 ± 7 nm, width: 56 ± 4 nm) with long non-contractile tails (length: 169 ± 9 nm, width: 10 ± 1 nm) and, as such, were classified as a member of the Drexlerviridae phage family [47, 48]. Such structural features are frequently associated with efficient host recognition and genome delivery, which may partly explain the favorable replication dynamics observed in this study. Similar morphologies have been widely reported among lytic Klebsiella phages [49, 50].

The infection dynamics of phages are largely determined by their adsorption efficiency, latent period, and burst size, which collectively define their capacity to attach to host cells, replicate, and produce progeny. Phages with rapid adsorption, short latent periods, and large burst sizes are generally more effective at suppressing bacterial populations, making them suitable candidates for therapeutic or biocontrol applications. These parameters should therefore be carefully measured when characterizing new isolates [51, 52]. In this study, UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 demonstrated rapid and efficient adsorption of about 95% to their host cells. High adsorption rates are characteristic of effective K. pneumoniae phages, as prompt and strong binding to bacterial receptors facilitates successful infection cycles, as reported by several studies. For instance, phage BUCT631 was reported to achieve over 90% adsorption within 15 min of exposure to K. pneumoniae host cells, reflecting similarly strong initial host interactions [50]. Other studies on Klebsiella phages have also documented high adsorption efficiencies within 10–20 min [53–55]. In contrast, Duarte et al. [56] reported that the K. pneumoniae phage KP1LMA reached only 60% adsorption after 60 min of exposure to host cells, illustrating that some phages exhibit substantially slower host-binding kinetics. Furthermore, the latent period observed for the three phages (~ 25 min) is well within the range reported for many K. pneumoniae phages. Several studies reported latent periods ranging between 5 and 30 min, with several phages exhibiting latent intervals around 20–25 min before progeny release, demonstrating comparable timing to the isolates in this study [50, 54, 57, 58]. While some phages may present exceptionally short or long latent periods depending on host physiology and experimental setup, the 25-minute interval exhibited by the three phages in this study suggests effective genome replication and assembly kinetics under the given assay conditions. The three phages UGKSKpnP1-P3 exhibited large burst sizes of approximately 311–340 PFU per infected cell, falling well within the range reported for K. pneumoniae phages. For instance, Baqer et al. [54] documented burst sizes of up to 350 PFU per cell in lytic Klebsiella phages. Similarly, studies by Peng et al. [49], Hans et al. [50], and Abdel-Razek et al. [59] reported high burst sizes ranging from 303 to 354 PFU per infected cell following the latent period. These findings collectively confirm that the burst sizes observed in the present study are indicative of strong lytic productivity. Although the variability in burst sizes among the three phages was minimal, such differences are known to be influenced by host cell physiology, phage genome packaging capacity, and experimental growth conditions, which may account for variations observed among isolates, even within the same species [54]. The replication dynamics observed for the three phages, characterized by rapid adsorption, short latent periods, and high burst sizes, underscore their strong lytic efficiency, and these attributes make them highly promising candidates for phage therapy and phage-mediated biocontrol applications against K. pneumoniae.

The impact of the environment on phage viability is a significant determinant of their potential use in phage therapy and other phage-mediated biocontrol applications, as environmental conditions affect phage infectivity and shelf life [7]. In this study, the three Drexlerviridae family phages (UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3) that infect K. pneumoniae exhibited robustness across a range of temperature and pH values exposure conditions. While tolerance towards ultraviolet (UV) exposure was limited. Therefore, in this study, phage viability remained stable from − 20 to 35 °C, consistent with other research showing thermal tolerance of K. pneumoniae phages within moderate temperature intervals of 4–50 °C without significant loss of infectivity [60]. Phage reductions observed between 40 and 65 °C, and complete inactivation by 70 °C, mirror previous reports where K. pneumoniae phages lost infectivity at higher temperatures, often above 60–80 °C [54, 60]. This is attributed to thermal denaturation of structural proteins. Temperature plays an important role in determining bacteriophage infectivity because it directly influences the stability of phage structural proteins. Exposure to elevated temperatures can destabilize the capsid and tail proteins, leading to conformational changes in receptor-binding components. Such alterations may reduce the ability of phages to recognize and attach to their specific bacterial receptors, ultimately resulting in reduced adsorption efficiency and loss of infectivity [60, 61]. Similarly, pH significantly influences phage activity, with conditions close to neutrality generally supporting greater stability [62]. This observation is consistent with the findings of our study, in which phage activity was highest at near-neutral pH. In contrast, a marked reduction in phage viability was observed under highly acidic conditions (pH 2) and at strongly alkaline pH levels. Notably, phage activity increased progressively as pH shifted from extreme acidic or alkaline conditions toward neutrality. This trend is comparable to reports on other K. pneumoniae phages, which show optimal lytic activity across a relatively broad pH range centered around neutrality [49, 55, 63]. The observed pH tolerance overlaps with physiologically relevant conditions, indicating that UGKSKpnP1-P3 are likely to retain activity in both environmental and host-associated settings. This breadth of pH stability enhances their practical utility, particularly for applications in which phages are exposed to fluctuating pH conditions, such as during gastrointestinal transit or in environmental biocontrol interventions. The marked reduction in phage viability at pH values outside this range likely reflects disruption of capsid protein structure and function under extreme acidity or alkalinity, where extensive protonation or deprotonation of ionizable amino acid residues alters electrostatic interactions responsible for maintaining capsid integrity and effective receptor binding, eventually compromising phage stability and infectivity [64, 65]. The rapid decline in phage titers under UV Light exposure observed in this study, with complete inactivation by 40 min, is comparable to other studies that used UV light to serve as a standard stress test for phage stability, given the potential for UV-mediated nucleic acid damage [53, 66–68]. Prolonged exposure to UV light damages nucleic acids and renders phage particles noninfectious, a phenomenon observed in studies that evaluated phage persistence under germicidal UV conditions [68–70]. The statistically similar UV response among UGKSKpnP1-P3 suggests conserved structural vulnerabilities within this phage group.

Whole-genome sequencing revealed that phages UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 possessed highly conserved genomic organization, as evidenced by their nearly identical genome sizes (~ 51.6 kb) and high intra-group nucleotide similarity (99.43% ANI). Such close relatedness strongly suggests that these phages represent very closely related variants originating from a common evolutionary lineage, a phenomenon frequently reported among environmentally co-circulating lytic phages infecting the same host species [71, 72]. These genomic findings are further corroborated by phenotypic data, as the phages exhibited statistically comparable responses to environmental factors influencing viability and replication, including temperature, pH, and UV light exposure, alongside similar adsorption efficiencies, latent periods, burst sizes, and host range profiles. Despite this high genomic similarity, minor differences in coding sequence numbers and functional annotations were observed, primarily involving hypothetical proteins. The predominance of genes with unknown function is typical of siphoviruses and highlights the extent of unexplored genetic diversity within Klebsiella phage populations [72]. Importantly, the absence of tRNA genes suggests a strong dependence on host translational machinery, consistent with relatively compact lytic genomes adapted to efficient replication rather than long-term host integration. Genome annotation confirmed the lack of integrases, recombinases, or other genetic markers associated with lysogeny, as well as the absence of virulence factors, toxin genes, and antimicrobial resistance determinants. This genomic profile strongly supports a strictly lytic lifestyle among the three phages, which is a fundamental prerequisite for phages intended for therapeutic or biocontrol applications [73, 74]. Whole-genome screening is therefore essential to minimize the risk of horizontal gene transfer and unintended enhancement of bacterial pathogenicity. Phylogenetic analyses based on whole-proteome comparisons, terminase large subunit sequences, and core gene sets consistently placed the three phages within the genus Webervirus of the family Drexlerviridae. Their clustering into a distinct and well-supported clade closely related to, but genetically separable from, ICTV-recognized species such as Klebsiella virus Call and Klebsiella virus KOX1 indicates that UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 represent a novel species within this genus. This conclusion is further reinforced by VIRIDIC analysis, which grouped UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 within the same species cluster (Cluster 45), indicating that they represent three highly similar isolates of a single species. The three genomes share ~ 99.43% intergenomic similarity, consistent with strain-level variation rather than distinct species. Notably, species cluster 45 is distinct from all previously reported Webervirus species clusters, indicating that UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 form an independent taxonomic grouping. In addition, the three phages shared at most 93.2% average nucleotide identity (ANI) with other analyzed Webervirus genomes. As proposed by Roux et al. [75], ANI values below the 95% species demarcation threshold support designation as a novel species. Together, the distinct clustering pattern and sub-threshold ANI values confirm that UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 represent a previously uncharacterized species within the genus Webervirus.

Limitation

One limitation of this study is that O-antigen typing of the Klebsiella pneumoniae strains was not performed. Although the isolated phages exhibited broad lytic activity across multiple capsular (K) types, the absence of O-antigen characterization limits definitive conclusions regarding the precise bacterial surface receptor(s) involved in phage adsorption. Since phages capable of infecting diverse capsular types may utilize alternative surface structures such as lipopolysaccharide components or outer membrane proteins, inclusion of O-antigen profiling would have provided additional insight into receptor specificity and host–phage interactions. Future studies incorporating detailed surface antigen characterization will help elucidate the molecular basis of host recognition and further refine the therapeutic potential of these phages.

Conclusion

The three phages UGKSKpnP1, UGKSKpnP2, and UGKSKpnP3 exhibited strong lytic activity against diverse carbapenem-resistant K. pneumoniae capsular pathotypes, high replication efficiency, and favorable infection kinetics, including rapid adsorption, short latent periods, and large burst sizes. Whole-genome analysis revealed highly conserved, strictly lytic genomes lacking lysogenic markers, virulence factors, or antibiotic resistance genes, underscoring their biosafety. Phylogenetic and VIRIDIC analyses placed the phages within the genus Webervirus while supporting their designation as a novel species. Collectively, these attributes highlight their promise as potential candidates for phage therapy and phage-mediated biocontrol against multidrug-resistant K. pneumoniae.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

This work was carried out in collaboration between all authors. KS, DKB, EW, and CKD conceptualized and designed the format for this study. KS, SM, ED, and IN carried out all the laboratory experiments. KS, EJ, SM, and CKD conducted data analysis. All authors drafted and managed manuscript revisions. All authors read and approved the final manuscript.

Funding

The authors declare that this research project was funded by the Africa Centre of Excellence in Materials, Product Development & Nanotechnology; MAPRONANO ACE, Grant No. P151847IDA credit 5797-UG, College of Engineering Design Art and Technology, Makerere University.

Data availability

Annotated phage genomes were deposited in the European Nucleotide Archive (ENA) under accession numbers: UGKSKpnP1 (OV877387), UGKSKpnP2 (OV877339), and UGKSKpnP3 (OV877766).

Declarations

Ethics and consent to participate

Ethical Approval No: MHREC1611 was granted by the Research and Ethics Committee for ethical review and approval, Mulago National Referral Hospital. The Research Ethics Committee of Mulago National Referral Hospital waived the need for informed consent to use already coded archived samples in this study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.