Isolation and characterization of lytic bacteriophages with therapeutic potential against multidrug resistant Klebsiella pneumoniae from Ethiopia

Abstract

Klebsiella pneumoniae is a common pathogen responsible for various infections, with multidrug-resistant (MDR) strains increasingly complicating treatment. Phage therapy has shown significant potential in treating difficult bacterial infections; however, research specifically focused on phages targeting this bacterium remains limited. This study aimed to isolate and characterize lytic bacteriophages that target multidrug-resistant K. pneumoniae. Phages were isolated from 66 environmental samples through spot assays conducted on 10 multidrug-resistant strains. Phages were isolated via spot assays and purified via streak plating. Characterization included PCR-based identification and classification, determination of the latent period, efficiency of plating, burst size, stability testing, and evaluation of in vitro bactericidal activity. From a total of 660 spots tested against 10 multidrug-resistant (MDR) isolates, 102 phages were successfully isolated. These phages demonstrated individual lytic activity ranging from 8% (4/46) to 63% (29/46). PCR-based classification of the 60 bacteriophages identified six distinct virulent phage genera, with Taipeivirus being the most prevalent at 18.3% (11/60) and Webervirus the least common at 10.0% (6 out of 60). Stability assessments of pH and temperature demonstrated optimal activity between pH 5 and 9 and at temperatures up to 50 °C. These results endorse phage therapy as a viable alternative for treating MDR and hypervirulent K. pneumoniae infections. The data offer critical insights into local bacteriophage diversity and underscore its potential for developing targeted therapeutic agents. Genome sequencing, in vivo studies, and clinical trials are required to validate the efficacy and safety of these phages.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-39153-8.

Introduction

Antibiotic resistance has become one of the world’s major public health concerns. Multidrug-resistant (MDR) Klebsiella pneumoniae (K. pneumoniae) poses a significant healthcare threat because of its broad antibiotic resistance and ability to cause severe infections¹. K. pneumoniae is a widely distributed gram-negative, nonmotile, encapsulated bacterium belonging to the Enterobacteriaceae family. It commonly colonizes the gastrointestinal tract, skin, and nasopharynx of humans and other mammals. K. pneumoniae is known to cause a range of infections, including urinary tract infections, abdominal infections, meningitis, suppurative liver abscesses, septicemia, and both nosocomial and community-acquired pneumonia. Its polysaccharide capsule is a key virulence factor that helps the bacterium evade host defenses and contributes to its pathogenicity².

The increasing incidence of drug resistance has necessitated the development of alternative therapeutic strategies beyond conventional antibiotics. The emergence of MDR and hypervirulent strains of K. pneumoniae poses a dual threat, significantly hindering the effectiveness of currently available antibiotics³. These situations underscore the pressing need for novel and innovative therapeutic approaches. In this setting, bacteriophages (phages) and their components emerge as viable alternatives to conventional antibacterial treatments⁴. To achieve this goal, isolating potent and effective bacteriophages is a requisite step. The targeted isolation and thorough characterization of phages capable of infecting and lysing multidrug-resistant bacteria have contributed to the development of precise and adaptable therapeutic options. Numerous studies have shown the potential of phages to effectively treat MDR K. pneumoniae infections both in vitro and in vivo^5–10.

Phage therapy exploits the natural bacterial host specificity and bactericidal activity of phages, providing targeted eradication of pathogens with minimal impact on the host microbiota. The renewed interest in phage therapy is driven by the unique ability of phages to coevolve with and overcome bacterial resistance mechanisms. Although phage research has experienced a global resurgence, data on the isolation, characterization, and therapeutic potential of lytic bacteriophages targeting MDR K. pneumoniae in Ethiopia remain limited. The isolation of potent lytic phages from various environmental and clinical sources is essential for developing effective phage cocktails that are specifically tailored to target MDR bacterial strains.

Phage diversity in Ethiopia remains largely unexplored and poorly documented. While phage research is gaining momentum across Africa, significant gaps persist in understanding the diversity and taxonomy of phages within Ethiopia, particularly those targeting multidrug-resistant bacteria. Given Ethiopia’s rich ecological variety, uncovering phage diversity in this region offers the potential to identify novel phages with unique therapeutic properties. Such efforts could complement existing antimicrobial stewardship programs and enhance clinical outcomes for patients infected with MDR K. pneumoniae. Therefore, this study aims to isolate lytic bacteriophages against these strains from various sources in Addis Ababa, characterize and classify them at the genus level via PCR, and assess their therapeutic potential in vitro.

The findings of this study highlight the rich diversity and therapeutic potential of lytic bacteriophages isolated from wastewater and sewage environments against multidrug-resistant K. pneumoniae. The observed variability in host range, phage stability, replication dynamics, and bactericidal efficacy underscores the importance of selecting and combining phages to broaden treatment coverage. The effectiveness of minimal phage cocktails in lysing a broad spectrum of clinical isolates further supports the potential of phage therapy as a targeted, adaptable, and eco-friendly alternative to conventional antibiotics for combating MDR K. pneumoniae infections. Additionally, PCR-based classification proved to be a rapid and sensitive tool for identifying and taxonomically differentiating virulent Klebsiella phages. To advance toward clinical application, comprehensive genomic characterization, in vivo validation, and rigorous clinical evaluation are essential. These steps will be critical for optimizing phage formulations, ensuring safety, and translating the potential of phage therapy into practical treatments for multidrug-resistant bacterial infections.

Materials and methods

Description of the study area

The study was conducted in selected areas of Addis Ababa from February 2024 to April 2025. The laboratory work was carried out at the Health Biotechnology Laboratory within the Biotechnology Research Centre of Addis Ababa University. Addis Ababa is the capital and largest city of Ethiopia. The population of Addis Ababa in 2024 was 5,704,000, a 4.45% increase from 2023. The city is located on a well-watered plateau surrounded by hills and mountains at an elevation of 2355 m above sea level. The city is located at latitude 9.005401°N and longitude 38.763611°E. A map of the study area and sampling sites appears in Fig. 1. On average, the city receives approximately 1165 mm of rainfall annually, which is primarily concentrated during the main rainy season from June to September, with July and August being the wettest months. The surrounding environment is ecologically diverse, featuring rich highland forests and a riverine ecosystem shaped by topography and climate¹¹.

Fig. 1.

Map of the study area and sampling locations. Sampling point coordinates were converted to Universal Transverse Mercator (UTM) coordinates for compatibility with the Addis Ababa city boundary shape file obtained from EthioNSDI GeoNode (2024). Mapping was performed using ArcGIS Desktop version 10.8 (Esri, Redlands, CA, USA).

Study design

A cross-sectional study was conducted from April 30, 2024, to February 2025 to isolate bacteriophages specific to K. pneumoniae from selected environmental sources. Using a convenient, non-probability purposive sampling approach, samples were collected from hospital sewage, wastewater, and soil at four major hospitals: Tikur Anbessa Specialized Hospital (TASH), Menelik II Referral Hospital, Yekatit 12 Hospital Medical College, and Zewditu Memorial Hospital were among the sampling sites, alongside fourteen wastewater-impacted river locations (Fig. 1). Additional details of the sampling sites are provided in Supplementary Table 4. These sites were chosen due to their high levels of organic waste and contamination, which create favorable conditions for the presence and proliferation of bacteriophages. Additionally, ten K. pneumoniae host bacterial isolates with high Multiple Antibiotic Resistance (MAR) indices were selected for bacteriophage isolation.

Sample collection, transportation, and processing

Sewage and wastewater samples were collected and processed following Shende et al.¹². Two sampling methods were employed: surface discrete sampling from the top layer and mixed sampling at depths of 25 cm and 45 cm below the surface. Forty milliliters of sample were collected in 50 mL Falcon tubes and transported in an ice box, then stored at 4 °C until processing. Samples were homogenized with 1 mL SM buffer (2 mM CaCl2, 10 mM MgSO4, 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 0.01% gelatin) by gentle inversion 10–15 times. Ten milliliters of homogenate were mixed with 15 mL of double-strength LB broth and incubated at 37 °C with shaking at 180 rpm for 18 h to amplify phages. The preparation was centrifuged at 6000 rpm for 20 min, followed by centrifugation of 2 mL supernatant at 10,000 rpm for 10 min. The final supernatant was filtered through a 0.22 μm syringe filter and stored at 4 °C for phage screening. Soil samples were collected as described by Mkwata et al.¹³. Twenty grams of soil were collected from depths of 5–25 cm, transported in sterile bags in an icebox, and stored at 4 °C. Ten grams were mixed with 50 mL double-strength LB broth and 100 µL SM buffer in a flask, then shaken 10–15 times and incubated at 37 °C and 180 rpm for 18 h. Ten milliliters were centrifuged at 6000 rpm for 20 min, and afterward, the same processing steps as for sewage samples were applied¹⁴.

Host strain and growth conditions

A total of 46 clinical bacterial isolates were included in this study. Ten K. pneumoniae isolates were used as hosts for bacteriophage isolation, while the remaining isolates were selected for host range assessment. A detailed list of bacterial isolates and their sources is provided in Supplementary Table 1. Isolates were cultured in 5 mL of tryptic soy broth (TSB) and incubated at 37 °C for 18 h. An established average OD6₀₀ value for 1 × 10⁸ CFU/mL of the isolation hosts is found in Supplementary Table 2. The bacterial strains used in this study were previously characterized and utilized in our earlier publication¹⁴.

Phage isolation and purification

Bacteriophages were isolated by spot assay and purified by streak plating as described by Abebe et al.¹⁴. Ten MDR K. pneumoniae strains served as propagation hosts. Cultures were grown in tryptic soy broth (TSB) to an optical density at 600 nm (OD₆₀₀) of 0.4–0.5. Bottom agar (TSA, 1.5% w/v) and soft agar (0.5–0.7% w/v) were prepared, with plates divided into labeled quadrants. Bacterial lawn was prepared by mixing 200 µL overnight culture, 40 µL each of 10 mM MgSO₄ and CaCl₂, and 3 mL of soft agar, then overlaid and solidified for 5 min. Twenty microliters of sample 20 were spotted, and the plates were incubated at 37 °C for 18 h to detect plaques/lysis. Individual plaques were streaked onto TSA plates, overlaid with soft agar containing bacterial lawns, and incubated at 37 °C for 18 h. Well-isolated plaques were picked, resuspended in SM buffer, vortexed, and incubated at 4 °C for 3 h. The suspension was centrifuged at 13,000 rpm for 10 min, filtered through a 0.22 μm syringe filter, and the resulting phage lysate was stored at 4 °C. This process was repeated three times to ensure purity.

Host range analysis

The host range of the bacteriophages was evaluated via spot tests¹⁴ on 46 multidrug-resistant clinical isolates. These included K. pneumoniae ATCC700603, E. coli ATCC25922, aeruginosa ATCC27853, Acinetobacter baumannii, Proteus mirabilis, K. oxytoca, and K. ozaenae. The isolated phages were tested against a variety of bacterial isolates obtained from different clinical sources. The complete list of bacterial strains used for host range analysis is provided in Supplementary Table 1. Briefly, 200 µL of each bacterial isolate at the logarithmic growth phase was mixed with 3 mL of soft agar and overlaid onto TSA plates. Five microliters of purified phage lysate were then spotted onto the bacterial lawn. The plates were incubated at 37 °C for 18 h. The presence of a lysis zone in the spotted area indicated positive phage activity against the host bacteria.

Phage titer determination

A double-layer agar overlay (DLA) method was employed to determine phage titers. Initially, the phage filtrate was serially diluted in SM buffer to achieve tenfold serial dilutions. For dilutions of 10⁻⁴, 10⁻⁶, 10⁻⁷, and 10⁻⁹, 100 µL of each dilution was mixed with 100 µL of bacterial suspension and allowed to adsorb for five minutes. The mixture was then combined with 3 mL of soft TSA (0.5% agar) and overlaid onto freshly prepared TSA bottom agar plates. The plates were incubated at 37 °C for 18 h. Visible plaques, ranging from 30 to 300 per plate, were counted to calculate the plaque-forming units per milliliter (PFU/mL)¹⁵. The experiment was conducted in triplicate to ensure accuracy.

Determination of the optimal multiplicity of infection

The optimal multiplicity of infection (MOI) was determined following a previously described method¹⁶. Phage and bacterial suspensions were mixed in varying ratios to achieve MOIs of 0.01, 0.1, 1, 10, and 100 in sterile 1.5 mL Eppendorf tubes. The preparations were incubated at 37 °C for 4 h, followed by centrifugation at 10,000 rpm for 1 min at 4 °C. The supernatant was then filtered through a 0.22 μm syringe filter, serially diluted, and analyzed via plaque assays. The dilution that yielded the highest average plaque-forming units (PFUs) was identified as the optimal MOI. The experiment was conducted in triplicate.

Determination of the efficiency of plating (EOP)

The efficiency of plating was determined to assess the infectivity of the isolated phages via a previously established protocol¹⁷. Bacterial strains that were susceptible during host range analysis were grown (to OD 6₀₀ = 0.4–0.5), and a plaque assay was conducted on 10^{− 6} fold diluted phage lysates. The preparation was incubated at 37 °C for 18 h, and the average number of plaques was counted. The efficiency of plating was determined by dividing the average PFU/mL of test bacteria by the average PFU/mL of host bacteria. EOP values were categorized as low productive(0.001 < EOP < 0.1), medium productive(0.1 < EOP < 0.5), highly productive(EOP ≥ 0.5), and inefficient (≤ 0.001)¹⁸.

Phage stability at different temperatures

The stability of the selected phages at different temperatures—25 °C, 37 °C, 40 °C, 50 °C, 60 °C, and 70 °C—was investigated as previously described by Abebe et al.¹⁴, with 4 °C used as a control. Briefly, 100 µL of phage preparation was added to 900 µL of normal saline in an Eppendorf tube for the assay. The tubes were then incubated at the corresponding temperatures for an hour, followed by slow cooling to room temperature. After incubation, dilution was performed, and the average phage titer at each temperature point was determined via a double-layer agar assay. The average percentage reduction in phage viability at each temperature point was determined in comparison to the control (4 °C).

Phage stability across various pH values

The stability of the phages under different pH values was also assessed at five different pH values, specifically at pH 3, 5, 7, 9, and 11. To conduct a pH assessment, 300 µL of a phage suspension was transferred to 2 mL of TSB solution adjusted to the intended pH. The preparation was then incubated at 37 °C for an hour. The untreated phage lysate suspended in SM buffer was used as a control for pH stability assessment. Three parallel experiments were conducted, and the average phage titer was determined via the double-layer agar technique. The average percentage reduction in phage viability at each pH value was determined in comparison with the untreated control.

Single-step growth curve experiment

Single-step growth curves of the selected phages were generated to determine the burst size and latency period. Each K. pneumoniae host strains were grown in TSB at 37 °C to mid-log phase (~ 1 × 10⁸ CFU/mL). Phage stocks with known titers were added to achieve an MOI of 1. For infection, 4 mL of the bacterial culture (~ 1 × 10⁸) was mixed with the corresponding volume of phage stock and incubated at 37 °C for 5 min to allow phage adsorption. The preparation was then centrifuged at 10,000 rpm for one minute to remove nonadsorbed phages. The liquid part was discarded, and the pellet was suspended in 4 mL of prewarmed fresh TSB and incubated at 37 °C with shaking. Samples of 100 µL were taken every five minutes for 70 min and immediately diluted in SM buffer. DLA was used to calculate the average PFU/mL. Burst size was estimated by dividing the average PFU/mL at the plateau by the initial average PFU/mL during the latent period¹⁹.

In vitro bacterial cell killing assay

The in vitro bactericidal effects of the 10 phages were assessed as previously described¹⁹. The phages were selected on the basis of their host range to ensure broad-spectrum activity. Briefly, bacterial suspensions (~ 1 × 10⁸ CFU/mL) in the exponential growth phase were mixed with an equal volume of phages at MOIs of 0.1, 1, 10, and 100. The mixtures were incubated at 37 °C with shaking, and optical density (OD₆₀₀ nm) measurements were taken every hour for 8 hours. K. pneumoniae cultures mixed with TSB broth were used as positive controls. The experiments were performed independently in triplicate. The average OD₆₀₀ nm value for each MOI was calculated, and the results were compared with those of the control. The experiment was conducted in triplicate.

Minimal phage cocktail design for complete bacterial coverage

A minimal phage cocktail capable of lysing the maximum number of tested bacterial isolates was designed using a computational approach based on a binary infection matrix derived from efficiency of plating (EOP) data. To identify the smallest combination of phages that collectively lyse all tested bacterial isolates, we applied a greedy algorithm implemented in the PhageCocktail R package (available at https://CRAN.R-project.org/package=PhageCocktail)²⁰. This package uses a heuristic approach to iteratively select phages that cover the largest number of bacterial isolates until complete coverage is achieved. After computational selection, the minimal cocktail was manually inspected by cross-referencing the algorithm’s output against the original EOP values and host range data to confirm complete coverage of all target isolates.

Molecular characterization of bacteriophages

Bacteriophage DNA extraction and quality assessment

Phage DNA was extracted via a previously established protocol²¹. The bacteriophages were selected on the basis of the host range and titer. Briefly, samples were first incubated with DNase and RNase at 37 °C for one hour to remove extracellular nucleic acids. To inactivate the enzymes, EDTA was added to achieve a final concentration of 20 mM. Next, the samples were treated with proteinase K (20 mg/mL) and incubated at 56 °C for 90 min. DNA purification was carried out via the EZ-10 Spin Column Bacterial Genomic DNA Miniprep Kit. The concentration and purity of the extracted DNA were measured with a Nanodrop 2000/2000 C Spectrophotometer (Thermo Scientific™, USA), and its integrity was confirmed via 1% agarose gel electrophoresis. Finally, the purified DNA was stored at -20 °C for future molecular analyses.

Molecular detection of phages via PCR

To identify and classify the isolated phages, specific primers (Table 1) targeting eleven genera of lytic Klebsiella phages (Drulisvirus, Przondovirus, Taipeivirus, Slopekvirus, Yonseivirus, Webervirus, Jiaodavirus, Sugarlandvirus, Marfavirus, Mydovirus, and Jedunavirus) were used as outlined in Kornienko et al.²². Detailed primer sequences and PCR conditions are provided in Supplementary Table 3.

Table 1.

List of primers used for the identification and classification of lytic Klebsiella phages.

Target	Primer name	Sequence 5’ − 3’	Product length (bp)
Internal virion protein D	Przond	FW: CGTACAACCAAGGKGAAGG RV: TCCGTGAACACATCRTACCC	463
Putative tail tube-associated baseplate	Taipei	FW: AGTTCTGAACACCAAAGGC RV: CCAACTCAGAGCCGTTCC	326
Capsid protein	Drulis	FW CGCTCCGTAACGATAAGATG RV: ACGCAGACCGATGTTGTAC	728
Putative major head subunit precursor	Weber	FW: CCTATGATGACGACTCAAAC RV: ATTGCCAGCCATCTTATCAG	422
Alpha-glucosyl-transferase	Jiaoda	FW: TGAACATCAAAGCAATTCGTG RV: AACCACAGAATGCCAGAATC	502
Putative replicative DNA helicase	Sugarl	FW: GATCTACCAAGCTGTCCAG RV: AGTCGTTGTTACTCGTTCC	307
Putative baseplate hub	Slopek	FW: TCAAAGAACAATACGAAGAGG RV: TTGCCATTGCTTCCAGAGAG	411
Virion structural protein	Jeduna	FW: ACTTCTATTGTCATGGCTGG RV: CACCTTACAGTTTAGCGTC	349
Hypothetical protein	Marfa	FW: GCACCTGAAGGCATTACCC RV: CCCATCAATAGAATAAAGCAC	252
Putative helicase	Mydo	FW: GATCGAAAAGAATGTCTGGG RV: TTGGTCTACGATAATATCACG	302
DNA polymerase I	Yonsei	FW: GCACGCCGACCTATCCCG RV: GCCACGGTCATTGATAAGC	482

FW: forward, RV: reverse, bp: base pair.

Agarose gel electrophoresis

The PCR products were analyzed via electrophoresis on a 1.5% (w/v) agarose gel (LA agarose, Jena Bioscience, Germany) and stained with an ethidium bromide solution (10 mg/mL) (HIMEDIA Laboratories, India). Five microliters of the PCR product were mixed with 3 µL of 6X loading dye and then loaded on the prepared agarose gel lane. Electrophoresis was conducted at 100 V for 1 h in 1X TAE buffer (40 mM Tris, 20 mM acetic acid, and 1 mM EDTA at pH 8.0). A 100 bp DNA Ladder (Thermo Scientific™ GeneRuler™, Waltham, Massachusetts, USA) was used to estimate the size of the PCR products. The resulting band patterns were visualized and documented via a UVITEC Cambridge gel documentation system (Cambridge, United Kingdom).

Quality control, statistical analysis, and data visualization

Quality control was performed for each experiment to ensure data accuracy and reliability. The sterility of the media was tested using bottom and soft agar controls without samples or host bacteria; growth on these indicated contamination. Filtrate controls, containing filtered samples spotted on agar without host bacteria, were used to detect bacterial contamination in phage lysates. Host controls involved overlaying soft agar with host bacteria, without sample spots, to confirm the absence of intrinsic lysis. Experimental plates were prepared by mixing host bacteria with soft agar and spotting samples to detect zones of lysis, indicating phage activity (Supplementary Fig. 1). Data were collected in triplicate and analyzed using ANOVA, with significance set at P ≤ 0.05. The Mann-Whitney U test was used to compare phage recovery from different sample types. Data visualization was performed using the ggplot2 package in R version 4.5.1.

Results

Phage isolation

Bacteriophages are recovered from diverse environmental sources, including wastewater, hospital sewage, and soil. A total of 66 samples were processed for phage recovery (Supplementary Table 4). From a total of 660 spots tested across the 10 MDR K. pneumoniae strains, 102 bacteriophages were successfully isolated (Supplementary Table 5). Phage isolates were named using the sample source and type, followed by the code of the host from which they were isolated. The maximum number of phages was isolated from the hosts TASP06 and TASP117, whereas the lowest number of phages was recovered from the host TASP101.

The majority of phages were isolated from wastewater and hospital sewage (Fig. 2). Analysis of phage recovery showed that homogenized wastewater and sewage samples yielded significantly more phages compared to those collected by the discrete sampling method. Mann-Whitney U tests indicated higher phage recovery in homogenized samples (mean = 2.73, n = 60) than in discrete samples (mean = 1.36, n = 30; W = 137, p = 0.0123). Soil samples had the lowest recovery (mean = 0.545, n = 12), with significant differences from both homogenized (W = 425, p < 0.001) and discrete samples (W = 332, p = 0.0251) (Fig. 2).

Fig. 2.

Distribution of phage recovery by sampling site and sample type.

Host range analysis

The host range varied considerably within the K. pneumoniae species, with individual phages lysing between 8.6% (4/46) and 63% (29/46) of the tested clinical isolates. Among the phages tested, 42.2% (43/102) demonstrated lytic activity against K. pneumoniae ATCC 700,603. Some phages exhibited cross-species lytic activity, notably against K. oxytoca and K. ozaenae, indicating a broader host range. Lytic effects on aeruginosa, E. coli, A. baumannii, and mirabilis were limited (Fig. 3). Cross-species activity was lowest against aeruginosa (5.8%, 6/102) and highest against E. coli (11.7%, 12/102) (Fig. 3). Interestingly, phages capable of infecting multiple species tended to have a lower plating efficiency against K. pneumoniae isolates. Representative images of spot assays and plaque morphology on bacterial lawns are shown in Fig. 4.

Fig. 3.

Host range summary heatmap. The X-axis shows phage isolates, and the Y-axis shows bacterial isolates. Red indicates lysis, while cyan indicates no lysis.

Fig. 4.

Illustrative images from the host-range analysis.

“A” illustrates the interaction of phages GGM-TASP09, KFS-TASP92, ZHS-TASP18, and TAM-TASP06 with host TASP-117. “B” depicts the interactions of the phages TTM-TASP76, TBD-TASP32, GKM-TASP92, and ZHS-TASP18 with the host TASP-92. “C” shows the interactions of the phages GKM-TASP92, ZHS-TASP18, GGM-TASP09, and AKS-TASP18 with the host TASP-22. “D” shows the interactions of the phages TTM-TASP18, ZHS-TASP18, GKM-TASP92, and KMF-TASP06 with the host TASP-32. “E” shows the front view of plate “D”, whereas “F” represents the host E. coli ATCC25922, which exhibited no susceptibility to the tested phages. The letters A, B, C, and D on the plates indicate different replicates of the experiment.

Phage titration

Phage titration was performed on the 60 phages exhibiting the broadest lytic host range against diverse clinical K. pneumoniae isolates. The phage ZHS-TASP18 demonstrated the highest average titer at 4.3 × 10¹¹ PFU/mL (Table 2), while the lowest titer of 1.5 × 10⁴ PFU/mL was recorded for phage HGM-TASP09.

Table 2.

Phage Titration results of selected phages.

Phage	Titer. PFU.mL-1	Phage	Titer in PFU.mL-1	Phage	Titer in PFU.mL-1
ADM-TASP06	4.2 × 104	BYM-TASP32	4.2 × 107	BYD-TASP101	2.6 × 108
KFM-TASP06	2.3 × 1010	MHM-TASP32	5.3 × 108	TTM-TASP101	4.5 × 107
KFM1-TASP06	3.6 × 108	TTD-TASP37	1.3 × 107	LAM-TASP101	5.1 × 108
BYM-TASP06	4.0 × 109	TTM-TASP37	2 0.0 × 107	TTM-TASP117	5.2 × 1010
YHM-TASP06	6.3 × 105	HGM-TASP37	5.1 × 106	TCD-TASP117	2.8 × 108
MHM-TASP06	1.7 × 108	GGM-TASP37	8.4 × 105	TSL-TASP117	2.3 × 107
AKM-TASP09	2.3 × 1010	BYM-TASP37	8.8 × 109	AKD-TASP117	1.3 × 105
GGM-TASP09	8.3 × 106	GKM-TASP37	2.1 × 107	AKM-TASP117	2.8 × 105
GKM-TASP09	5.6 × 104	ZHM-TASP37	7.8 × 108	AKM1-TASP117	3.4 × 106
ZHS-TASP18	4.3 × 1011	TTM-TASP50	5.2 × 107	HGD-TASP117	7.4 × 107
TAD-TASP18	2.0 × 104	AKM-TASP50	3.6 × 108	GKM-TASP117	5.5 × 109
KFD-TASP18	5.6 × 108	TTM-TASP76	5.4 × 104	JMM-TASP117	4.0 × 107
AKS-TASP18	1.3 × 1010	GGM-TASP76	6.2 × 108	GGM-TASP117	2.6 × 108
YHD-TASP18	2.4 × 105	GKM-TASP92	9.1 × 108	HBM-TASP117	5.8 × 105
TTM-TASP32	5.2 × 105	TTM-TASP92	3.4 × 108	BYD-TASP92	6.1 × 106
TBD-TASP32	6.2 × 108	ADD-TASP92	3.3 × 109	MHD-TASP32	3.7 × 105
ADM-TASP32	2.5 × 1010	ADM-TASP92	5.1 × 105	HGM-TASP09	1.5 × 104
KFM-TASP32	5.7 × 104	GGM-TASP92	2.8 × 105	GKM-TASP18	2.6 × 106
GWM-TASP32	7.3 × 109	GGD-TASP101	1.4 × 107	TTD-TASP92	7.4 × 107
KID-TASP32	3.6 × 106	KIM-TASP101	4.7 × 106	GWM-TASP117	3.1 × 107

Optimal multiplicity of infection

The optimal multiplicity of infection (MOI) was investigated for ten selected phages, revealing MOI values ranging from 0.01 to 100 (Table 3). The phage AKM-TASP09 presented the lowest MOI of 0.01 with the highest average titer of 5.1 × 10¹⁰, whereas the highest MOI observed was 100 for the phages MHM-TASP32 and LAM-TASP101, with the highest average titers of 6.8 × 10⁸ and 3.8 × 10^8, respectively (Table 3).

Table 3.

Optimal multiplicity of infection.

1	Phage	MOI	Average highest titer
	KFM-TASP06	1	9.3 × 109
2	AKM-TASP09	0.01	1.2 × 1010
3	ZHS-TASP18	0.1	6.3 × 1011
4	AKM-TASP50	0.1	7.4 × 108
5	GKM-TASP92	10	1.1 × 109
6	TTM-TASP117	0.1	4.6 × 1010
7	MHM-TASP32	100	6.8 × 108
8	BYM-TASP37	10	1.2 × 1010
9	LAM-TASP101	100	3.6 × 108
10	GGM-TASP76	0.1	5.8 × 108

Phage efficiency of plating (EOP)

Phage efficiency was evaluated in 42 bacterial isolates that were susceptible during host range analysis. The EOP results were categorized into five groups: highly productive, moderately productive, low productive, inefficient, and not applicable (no lysis was observed during the host-range analysis). Phage‒host interactions exhibited considerable variability across bacterial isolates, with most phages demonstrating a spectrum of lytic efficiencies categorized as highly productive, moderately productive, or low productive (Fig. 5). Phage KFM-TASP06 demonstrated seven highly productive interactions, eleven with moderate productivity and four with low productivity. In comparison, the phages MHM-TASP32 and GGM-TASP76 presented four instances of high lytic productivity, accompanied by varying moderate and low productivity. Inefficient plating was rare, occurring only once with phage BYM-TASP37 (Fig. 5). Notably, K. pneumoniae ATCC 700,603 exhibited multiple highly and moderately productive interactions, whereas E. coli ATCC 25,922 and aeruginosa ATCC 27,853 exhibited low productivity (Fig. 5). Additionally, K. pneumoniae isolates such as TASP126 and TASP109 displayed strong susceptibility to phage infection.

Fig. 5.

Phage efficiency of plating heatmap. EOP values were categorized as low productive (0.001 < EOP < 0.1), moderately productive (0.1 ≤ EOP < 0.5), highly productive (EOP ≥ 0.5), and inefficient (EOP ≤ 0.001). The dark blue color in the heatmap represents “NA” (not applicable), indicating that the phage did not lyse the corresponding bacterial isolate. These experiments were conducted in triplicate.

Molecular characterization of phages

PCR-based identification and classification of phages

Phages were classified at the genus level using genus-specific primers. These included internal virion protein D for Przondovirus; putative tail tube-associated baseplate protein for Taipeivirus; capsid protein for Drulisvirus; major head subunit precursor for Webervirus; alpha-glucosyl transferase for Jiaodavirus; replicative DNA helicase for Sugarlandvirus; baseplate hub for Slopekvirus; virion structural protein for Jedunavirus; hypothetical protein for Marfavirus; helicase for Mydovirus; and DNA polymerase I for Yonseivirus (Fig. 6). The presence of distinct, appropriately sized markers confirmed the genus-level classification of the phage isolates.

Fig. 6.

Agarose gel electrophoresis analysis of the PCR products of K. pneumoniae phages. Lane M: DNA ladder (100 bp), lane NC: negative control, lanes with a number indicate sample lanes representing the PCR products of the corresponding genus-specific genes. “A” Putative tail tube-associated baseplate protein-encoding gene of the genus Taipeivirus (326 bp), “B” putative base plate gene of the genus Slopekvirus (411 bp), “C” capsid protein-encoding gene of the genus Drulisvirus (728 bp), “D” putative major head subunit precursor gene of the genus Webervirus (422 bp), “E” putative helicase gene of the genus Mydovirus (302 bp), “F” internal virion protein D of the genus Przondovirus (463 bp).

Among the 60 PCR-screened phages, 87.6% (52/60) were classified into six distinct genera. Taipeivirus was the most prevalent genus among the PCR-classified phages, whereas Webervirus was the least represented (Fig. 7). The comprehensive phage genera profile can be found in Supplementary Table 4.

Fig. 7.

Sources and genus-level taxonomic distributions of bacteriophages identified via PCR-based classification.

Single-step growth curve experiment

The latent period, defined as the time interval between phage adsorption and cell lysis, ranged from 15 to 35 min (Fig. 8). Phages TTM-TASP117 and BYM-TASP37 exhibited the shortest mean latent periods of 15 min. In contrast, phage LAM-TASP101 presented the longest latent period (Fig. 8). Burst size, which represents the number of progeny phages released per infected host cell, varied considerably among the phages (Fig. 8). The greatest burst size was observed for phage AKM-TASP50, with an average of approximately 323 PFU per infected cell. Other burst sizes included ZHS-TASP18 (318 PFU), LAM-TASP101 (218 PFU), and AKM-TASP09 (228 PFU). Phages with shorter latent periods, TTM-TASP117 and MHM-TASP32, produced average burst sizes of approximately 145 and 135 PFU, respectively.

Fig. 8.

Single-step growth curves of seven phages showing replication dynamics over time. The curves depict the latent, rise, and plateau phases for each phage. Data points represent the mean phage titer (PFU/mL) from triplicate experiments, with error bars indicating standard deviation.

Phage stability across various pH values

All ten tested phages demonstrated the highest viability within the pH range of 5–9, as shown in Fig. 9. At extreme pH values, phage viability was markedly reduced. Following a 60-minute incubation at acidic pH 3, the phage titer decreased by 75%. Similarly, exposure to alkaline pH 11 for 60 min resulted in a 50% reduction in phage viability (Fig. 9). Two-way ANOVA revealed that pH had a statistically significant effect on phage stability (p = 0.01), whereas phage type alone did not have a significant effect. Additionally, no statistically significant differences were observed between the different phage types.

Fig. 9.

Phage stability across a pH gradient. (A) Viability of phages across different pH levels, (B) percentage reduction in phage titer compared to the control. Data represent means ± standard deviations from three independent experiments.

Phage stability at different temperatures

The phage titers remained stable at 25 °C, 37 °C, and 40 °C, with no observed reduction (Fig. 10). However, exposure to higher temperatures caused a decline in phage viability. After 60 min at 50 °C, the phage titer decreased by 25%. This reduction increased to 50% at 60 °C. At 70 °C, only 25% of the phage titer was retained for most phages. Notably, the phages AKM-TASP09 and AKM-TASP50 did not survive after 60 min at 70 °C. In contrast, LAM-TASP101 exhibited the highest thermal stability among the tested phages, maintaining a relatively high titer (Fig. 10).

Fig. 10.

Phage stability across a temperature gradient. (A) Viability of phages at different temperatures, (B) percentage reduction in phage titer compared to the control (4 °C). Data represent means ± standard deviations from three independent experiments.

In vitro bacteria killing assay

The absorbance of the control culture (without phages) increased continuously (Fig. 11). In contrast, the absorbance of cultures treated with phages decreased in an MOI-dependent manner. MOIs of 10 and 100 markedly inhibited bacterial growth, whereas an MOI of 0.1 was associated with an increase in absorbance (Fig. 11).

Fig. 11.

MOI-dependent in vitro bactericidal effects of phages. Data represent means ± standard deviations from three independent experiments.

Minimal phage cocktail design for complete bacterial coverage

A minimal phage cocktail achieving complete coverage of 42 bacterial isolates was designed computationally using a greedy algorithm from the PhageCocktail R package, based on a binary infection matrix derived from EOP data (Fig. 12). The optimal four-phage cocktail (TTM-TASP117, GKM-TASP92, ZHS-TASP18, BYM-TASP37) was validated against original EOP and host range profiles, confirming alignment with individual phage susceptibilities and ensuring each isolate was lysed by at least one phage.

Fig. 12.

Comprehensive heatmap of minimal phage cocktail lysis patterns.

Discussion

In the current study, we isolated 102 bacteriophages from a total of 66 environmental samples processed against 10 MDR K. pneumoniae host strains. These phages demonstrated lytic activity against MDR clinical isolates, highlighting their therapeutic potential. Furthermore, the use of multidrug-resistant clinical strains as propagation hosts during isolation renders our phage collection a valuable resource for developing phage therapy against MDR K. pneumoniae. The majority of phages, 84.3% (86/102), were recovered from wastewater and sewage samples, indicating that these environments are rich reservoirs for phages that target K. pneumoniae. This finding aligns with prior studies showing that wastewater and sewage serve as major reservoirs of bacteriophages due to their high bacterial density and diversity²³. In contrast, the soil samples presented relatively low recoveries. This finding aligns with previous studies showing that soil environments harbor fewer phages compared to aquatic or sewage sources²⁴. This lower abundance may be attributed to environmental constraints that limit phage survival or accessibility. Additionally, this can be explained by the limited survival and low abundance of K. pneumoniae in soil compared with nutrient-rich environments such as sewage and wastewater²⁵. As an enteric bacterium, K. pneumoniae primarily thrives in animal intestines and aquatic habitats, leading to fewer available hosts and a lower abundance of phages in the soil²⁶.

In the present study, the host range of 102 phages was assessed against 46 clinical bacterial isolates, which revealed a wide range of lytic activity. Individual phages lysed 8.7% (4/46) to 63% (29/46) of the tested clinical isolates. Notably, 42.2% (42/102) of the phages lysed K. pneumoniae ATCC700603, which is consistent with reports of variable host ranges within K. pneumoniae due to its genetic diversity and surface receptor variability²⁷. Earlier studies reported that K. pneumoniae phages can achieve lysis rates of up to 68% against clinical isolates. For example, phage KPAФ1 lyses approximately 26.15% of MDR K. pneumoniae strains, with phage cocktails increasing this coverage to approximately 32.30%²⁸. Another notable phage, M198, lyses approximately 60% of the strains tested (62 out of 103), resulting in high plating efficiency for nearly 19.4% of these isolates and intermediate sensitivity to other strains²⁹. Some phages, such as vB_KleM_KB2, demonstrated strong lytic activity and effectively inhibited bacterial growth at low multiplicities of infection. In contrast, many phages have much narrower host ranges, lysing only a small fraction of clinical isolates as low as 2%³⁰.

In the present study, some phages exhibited cross-species lytic activity, primarily against K. oxytoca and K. ozaenae. These findings support the idea that polyvalent phages can infect multiple species within a genus via shared surface receptors. However, their activity against more distant genera, such as aeruginosa, A. baumannii, and mirabilis, was limited. Only 5.8% (6 out of 102) of the phages were active against aeruginosa, and most displayed low EOP. This finding suggests that the host range of phages becomes more restricted as the phylogenetic distance between bacterial hosts increases³¹. In the present study, we observed that phages exhibiting cross-genus infectivity presented lower lytic efficacy against K. pneumoniae, suggesting that they are not specifically adapted to this species. The reduced lytic efficacy of cross-genus phages against K. pneumoniae can be attributed to the diversity of bacterial receptors, as supported by previous studies^32–34.

Our study demonstrated complete lysis coverage of 42 bacterial isolates via a minimal cocktail of four phages: TTM-TASP117, GKM-TASP92, ZHS-TASP18, and BYM-TASP37. This finding supports previous research indicating that phage combinations can expand the host range and improve efficacy³⁵. These findings confirm that carefully selected phage cocktails can overcome the limited specificity of individual phages³⁶. Furthermore, the use of fewer phages simplifies production, lowers costs, and reduces the risk of resistance by targeting bacteria through multiple mechanisms. These results highlight the potential of phage cocktails in treating MDR infections. The phage titers in the present study varied significantly, ranging from 1.5 × 10⁴ for phage HGM-TASP09 to 4.3 × 10¹¹ PFU/mL for phage ZHS-TASP18. This variability is consistent with previous reports indicating that phage concentrations depend on the nature of the isolate and the methods used. For example, a study conducted on K. pneumoniae phage reported phage titers ranging broadly from 2 × 10^4 to 2.3 × 10¹¹ PFU/mL, demonstrating significant variability depending on the isolate and propagation conditions²⁹.

Potent lytic phages isolated in earlier studies often exceed 10⁹ PFU/mL, making them suitable for therapeutic applications. For example, phage hvKpP3 achieved a titer of 3.12 × 10¹⁰ PFU/mL, demonstrating strong efficacy against MDR strains of K. pneumoniae³⁷. The titers observed in this study for most phages are consistent with other reports of potent lytic Klebsiella phages, which achieve titers exceeding 10^⁹ PFU/mL³⁸. In contrast, lower titers, such as those of phage HGM-TASP09 in our study, reflect variations in replication and host interactions. These findings highlight the importance of optimizing phage selection and propagation to increase the effectiveness of phage therapy.

In this study, the MOIs of the ten K. pneumoniae phages varied from 0.01 to 100, indicating diverse phage‒host dynamics. The phage AKM-TASP09 exhibited a low MOI of 0.01, demonstrating its strong infectivity even at low ratios. In contrast, the phage LAM-TASP101 required a high MOI of 100, suggesting that higher doses are necessary for effective replication. These findings align with previous studies showing that optimal MOIs vary on the basis of phage and host characteristics in Klebsiella phages³⁹. For example, potent K. pneumoniae phages such as ΦK2044 and JKP2 have been shown to achieve high titers and effectively lyse bacteria. This occurs at low MOIs ranging from 0.001 to 0.1^35,40. Conversely, some research suggests that higher MOIs may be necessary for optimal phage amplification, depending on the specific phage‒host system⁴¹. Furthermore, recent studies indicate that lower MOIs can enhance phage replication and extend bacterial suppression by preventing the rapid depletion of host cells⁴². This highlights the importance of tailoring phage therapy to the specific interactions between phages and their hosts to maximize efficacy and minimize resistance when treating K. pneumoniae infections.

The one-step growth curve analysis of the ten phages in our study revealed latent periods ranging from 15 to 35 min, with burst sizes varying between 135 and 325 PFU per infected cell. These findings align with previous reports indicating that most Klebsiella phages have latent periods ranging from 5 to 50 min. Their burst sizes typically range between 30 and over 2,000 PFU per cell⁴³. For example, studies on K. pneumoniae phages such as vB_kpnM_17 − 11 and vB_KpnS_SXFY507 recorded latent periods of approximately 20–35 min and burst sizes of approximately 250–300 PFU per cell⁴⁴, which are consistent with our results. Similarly, Lin et al.⁴⁵ reported latent periods of 10–30 min and burst sizes ranging from 100 to 300 PFU per infected cell.

The observed inverse relationship between the latent period and burst size, where phages with shorter latent periods tend to have lower burst sizes, illustrates the classical trade-off in phage replication dynamics⁴⁶. The observed trade-off between the latent period and burst size is consistent with established principles of lytic phage biology⁴⁷. This pattern was evident in the phages TTM-TASP117 and MHM-TASP32, which demonstrated rapid replication cycles but produced fewer progeny. In contrast, phages such as AKM-TASP50 and ZHS-TASP18 presented relatively short latent periods alongside high burst sizes, suggesting efficient replication and assembly mechanisms that may offer advantages in therapeutic or biocontrol applications.

In our study, assessment of the stability of selected phages across various pH values revealed optimal viability within a pH range of 5–9. A significant reduction in phage titer was observed under acidic (pH 3) and alkaline (pH 11) conditions. For example, after 60 min at pH 3, certain phages presented a decrease in titer of up to 75%, whereas at pH 11, the reduction was approximately 50%. These observations align with earlier reports on Klebsiella and other gram-negative phages, which generally demonstrate stability in neutral to mildly acidic or alkaline pH ranges but lose viability under extreme pH conditions⁴⁸. Earlier studies on the Klebsiella phage vB_Kox_ZX8 have demonstrated stability across a pH range of 3–11, with some phages maintaining activity at pH 3, which is consistent with our observations. The variation in titer reduction among phages indicates differing levels of pH tolerance⁴⁹. Other studies conducted on K. pneumoniae phages demonstrated stable lytic activity across a broad pH range^50,51, supporting the potential of phages as antibacterial agents under diverse environmental conditions. Conversely, some phages, such as vB_KleM_KB2, exhibited a narrow pH tolerance and were stable only near neutral pH, which contrasts with the broader pH stability observed in our study and others⁵². The difference might arise from phage-specific structural or genomic features.

In the present study, phage ZHS-TASP18 exhibited greater stability at lower pH values than the other isolates did, maintaining its viability under acidic conditions that typically inactivate most phages. This acid tolerance is likely due to structural adaptations of phage proteins that protect against proton-induced damage⁵³. In contrast, phage TTM-TASP117 remained stable at pH 11, showing an adaptation to alkaline environments commonly found in sewage or industrial settings^53,54. These differences highlight the impact of environmental pressures on phage stability and have significant implications for their use in therapy and biocontrol across varying pH conditions⁵⁵.

The statistically significant effect of pH on phage stability observed in our study highlights the critical role of pH in affecting phage stability. This finding is consistent with other reports that emphasize pH as a key factor affecting phage infectivity and survival⁵¹. Moreover, the lack of significant differences among the phage types in our study suggests that these phages may have similar biochemical or structural features that influence their tolerance to different pH levels. Although we do not have direct evidence, this shared similarity likely explains why all the phages respond similarly to changes in pH. However, the exact mechanisms behind this pH tolerance are not yet fully understood. This highlights the need for further investigations to clarify the underlying biochemical or structural factors that govern phage stability and function under various pH conditions. Understanding phage stability is essential for effective therapy. Phages that are sensitive to extreme pH levels may require protective formulations, such as encapsulation or lyophilization, to preserve their infectivity during storage and delivery in the body⁵⁶. This is especially important for treating infections, as phages must navigate diverse pH conditions within the body.

The thermal stability of the ten phages examined in this study was assessed, revealing that the phage titer remained consistent at 25 °C, 37 °C, and 40 °C but decreased progressively at higher temperatures. There was a 25% reduction in titer at 50 °C and a 50% reduction at 60 °C and 70 °C. Compared with the control, seven even phages maintained 25% of their titer. Notably, the phages AKM-TASP09 and AKM-TASP50 were completely inactivated after 60 min of exposure at 70 °C, indicating extreme thermal sensitivity. These findings are consistent with previous reports indicating that many bacteriophages maintain infectivity at moderate temperatures but lose infectivity when the temperature increases beyond 40–50 °C. Amhaid et al. reported that Listeria phages p100 and A511 remained stable at 45 °C; however, P100 was notably sensitive to relatively high temperatures at 65 °C. In contrast, phage A511 demonstrated greater thermal stability under the same conditions⁵⁷.

Similar thermal profiles have been reported for other Klebsiella phages. For example, Klebsiella phage vB_kleM_KB2 maintained 90% viability after 30 min at 60 °C but was almost completely inactivated at 70 °C, retaining only 0.05% viability. This finding supports the loss of infectivity observed in AKM-TASP09 and AKM-TASP50 at 70 °C⁵². Other studies assessing the thermal stability of Klebsiella phages support these findings⁵⁸. For example, phage MKP-1 exhibited high stability over a wide temperature range (10–60 °C), maintaining 25% activity at 50 °C. However, it showed significant inactivation beyond this point, which aligns with the moderate thermal sensitivity observed in the current study.

For example, two K. pneumoniae phages, KpTDp1 and KkTDp2, demonstrated remarkable stability at temperatures up to 70 °C for 60 min, retaining their lytic activity. In contrast, several other Klebsiella phages, such as vB_KpnA_SCNJ1-Z, vB_KpnS_SCNJ1-C, and vB_KpnM_SCNJ1-Y, exhibited stability only up to 60 °C and were completely inactivated at 70 °C^59–61. Most studies characterizing lytic phages against K. pneumoniae and other gram-negative bacteria have shown that 70 °C is lethal for all tested phages⁵⁹. The irreversible loss of infectivity at 70 °C is likely due to the denaturation of essential phage proteins, a phenomenon observed in many mesophilic bacteriophages^59–61. From a practical standpoint, the thermal stability of the phages isolated in this study is highly advantageous for phage therapy applications. Since the average human body temperature is approximately 37 °C and most clinical and environmental conditions rarely exceed 40 °C, these phages have stable infectivity under in vitro conditions that simulate typical physiological environments; however, their effectiveness in vivo remains to be experimentally confirmed.

In the present study, the in vitro bactericidal effects of ten phages were evaluated against MDR K. pneumoniae. Previous studies reported that phages can significantly reduce MDK K. pneumoniae in vitro, particularly at MOIs of 1 or greater^62–64, supporting their role as potential antibacterial agents against MDR K. pneumoniae. Similarly, the phages used in the present study effectively inhibited the growth of MDR K. pneumoniae in an MOI-dependent manner. These findings highlight the therapeutic potential of these phages as alternatives to antibiotics.

In this study, 60 phages were classified via genus-specific PCR primers that target conserved genes of virulent Klebsiella phages²². Six genera, namely, Taipeivirus, Przondovirus, Slopekvirus, Mydovirus, Drulisvirus, and Webervirus, were detected. The current study did not detect five expected phage genera. This selective detection observed in this study likely reflects the high specificity of the primers, which target conserved genes in virulent phages but may overlook phages with divergent sequences or those lacking target genes. Additionally, local bacterial hosts may influence regional phage diversity. The presence of temperate phages and mutations in primer-binding sites might have contributed to the lack of detection. Studies have shown that phage types are often limited to infecting specific strains of bacteria, making it challenging to develop universal PCR detection methods⁶⁵. Similarly, PCR-based typing for virulent K. pneumoniae phages has demonstrated limited coverage due to sequence divergence among phage genera and the presence of novel or uncharacterized phages²². Given the limitations of PCR-based methods, whole genome sequencing is crucial for accurately characterizing phage diversity. Whole genome sequencing enables identification of both known and novel phages, including those with mutated or divergent sequences missed by PCR, and provides detailed insights into their genetic makeup, host range, and virulence factors.

The PCR results revealed that the highest diversity and abundance of phages were observed in the wastewater samples, which contained all six identified genera. Notably, Slopekvirus and Mydovirus were particularly prevalent. Sewage samples also presented a greater presence of Taipeivirus and Przondovirus phages. The high abundance of phages in sewage and wastewater environments can be attributed to several ecological factors. These settings are rich in organic matter and support high densities of diverse bacterial hosts, such as K. pneumoniae, creating favorable conditions for phage proliferation⁶⁶. Additionally, the continuous influx of bacteria from human and animal sources provides a dynamic reservoir for both phage and bacterial evolution, likely contributing to the increased phage diversity and abundance observed in these environments. Previous studies have similarly indicated that urban wastewater and sewage serve as hotspots for isolating lytic phages against clinically relevant bacteria, including K. pneumoniae^67,68. The dominance of the Taipeivirus and Przondovirus genera among the PCR-detected phages in the current study agrees with previous reports identifying these genera as common and effective lytic phages against K. pneumoniae⁶⁹. The relatively high prevalence of these phages in our study may suggest ecological adaptation of phages to the K. pneumoniae strains present in the sampled environments. This observation is consistent with the concept that phage populations can coevolve with their bacterial hosts, influencing the composition and structure of phage communities within specific geographic regions⁷⁰. However, further studies are needed to confirm the dynamics of this coevolution in natural settings.

In this study, most isolated bacteriophages in this study effectively inhibited the targeted MDR K. pneumoniae strains. However, further investigations are needed to evaluate their activity against a broader spectrum of pathogens, assess their stability under various environmental conditions, and comprehensively characterize their genomic features. Ongoing studies are aimed on whole genome sequencing to elucidate phage genetic diversity, identify potential virulence or resistance factors, and ensure the safety and efficacy of phage candidates for therapeutic or bio preservation applications.

Conclusion

The findings of this study highlight the rich diversity and therapeutic potential of lytic bacteriophages isolated from wastewater and sewage environments against multidrug-resistant K. pneumoniae. The observed variability in host range, phage stability, replication dynamics, and bactericidal efficacy underscores the importance of selecting and combining phages to broaden treatment coverage. The effectiveness of minimal phage cocktails in lysing a broad spectrum of clinical isolates further supports the potential of phage therapy as a targeted, adaptable, and eco-friendly alternative to conventional antibiotics for combating MDR K. pneumoniae infections. Additionally, PCR-based classification proved a rapid, sensitive tool for identifying and taxonomically differentiating virulent Klebsiella phages across different genera. The presence of various genera reflects the ecological diversity of the study area and suggests its potential as a source for isolating lytic phages. To advance toward clinical application, comprehensive genomic characterization, in vivo validation, and rigorous clinical evaluation are essential. These steps will be critical for optimizing phage formulations, ensuring safety, and translating the potential of phage therapy into practical treatments for multidrug-resistant bacterial infections.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

AAA: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing -original draft, Writing - review & editing AGB: Conceptualization, Validation, Writing -review & editingTST: Conceptualization, Funding acquisition, Project administration, Resources, Validation, Writing – review & editing.

Data availability

All the data generated or analyzed during this study are included in this article and its supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.

Ethical clearance

The study proposal was reviewed and approved by the Research Ethics Committee at the Institute of Biotechnology, Addis Ababa University (IoB/431/2016/2024). All methods were carried out in accordance with relevant ethical guidelines and regulations.