Unveiling Novel Lytic Bacteriophages as Natural Biocontrol Agents Against Multidrug-Resistant Escherichia coli: Isolation, Characterization, and In vitro Application

TL;DR

This study isolates and characterizes four new bacteriophages that effectively target multidrug-resistant Escherichia coli, showing potential for biocontrol applications.

Contribution

The discovery of novel lytic bacteriophages with strong activity against MDR-E. coli and their classification as new Vectrevirus members.

Findings

Phage Sem4 showed stronger lytic activity than a phage cocktail, fully suppressing MDR-E. coli growth.

The phages remained stable for up to a year at 4°C and effectively reduced bacterial counts in tap water.

Genomic analysis revealed novel Vectrevirus members with no genes linked to antibiotic resistance or lysogeny.

Abstract

The alarming findings presented in the latest WHO report on the global antimicrobial resistance crisis have redirected scientific attention toward phage-based approaches as a renewed line of defense against multidrug-resistant (MDR) bacteria. In this study, four bacteriophages infecting a MDR-Escherichia coli strain were isolated from water sources and subjected to detailed phenotypic and genomic characterization. All phages efficiently inhibited MDR-E. coli at MOIs of 0.1 and 0.01, showing high stability across a broad temperature (4–65 °C) and pH (4–10) range. TEM analysis revealed that all phages exhibited a podovirus-morphotype. At 4 °C, titers remained stable for 6 months, with only a 1–2 log reduction -over a year. Notably, phage Sem4 exhibited markedly stronger lytic activity than the phage cocktail, fully suppressing bacterial growth. In tap water, phage Sem4 treatment reduced bacterial counts from 10⁷ to 7 × 10⁴ CFU/mL within 8 h, with no detectable colonies at 24–48 h. Genomic analysis showed that these phages possess linear dsDNA genomes of 44,244–45,205 bp, with ~ 45% GC content. Phylogenetic and comparative analyses classified them as novel Vectrevirus members within the Molineuxvirinae subfamily of the Autographiviridae family, sharing less than 95% intergenomic similarity with known Vectrevirus species. No genes associated with antibiotic resistance, toxins, or lysogeny were detected. These findings underscore the potential of - phages as a promising candidate for the development of next-generation biocontrol strategies, especially marking the efficiency of Sem4 in water sanitation systems, paving the way for sustainable and targeted interventions against MDR bacterial contamination.

The online version contains supplementary material available at 10.1007/s12560-026-09684-4.

Figures

In vitro characterization of phages Sem1, Sem2, Sem3, and Sem4 (*i*–*iv*, respectively). A Representative plaque morphology of the phages formed on bacterial lawns of their respective hosts. B Transmission electron micrographs of negatively stained phage particles (2% [w/v] uranyl acetate), displaying typical podovirus morphology; scale bars represent 0.1 μm. C OSGCs of the phages on MDR-Ec. Y-axis denotes log₁₀-transformed phage titers at distinct time points. Error bars represent standard deviations from three independent replicates (= 3). D The effect of different MOIs of the phages on the

Stability and antibacterial activity assays of phages Sem1, Sem2, Sem3, and Sem4, and biocontrol performance of Sem4 in tap water. A–C Stability profiles of the phages under different pH conditions (A), temperatures (B), and long-term storage (C; percentages indicate the relative change in phage titers at the end of the incubation period) (= 3). D Minimum inhibitory concentration assay results for the four phages across two growth media (TSB and MHB) and four phage concentrations (top to bottom: 5 × 10^7^, 5 × 10^5^, 5 × 10^3^, and 5 × 10^1^ PFU/mL), measured via OD_600_ absorbance at the end

Linear genomic representation of the phages. CARD analysis results are indicated in red. GC content, GC skew, and ORFs are shown in distinct colors. A ORFs identified in phage Sem1 are shown in fuchsia. B ORFs identified in phage Sem2 are shown in purple. C ORFs identified in phage Sem3 are shown in fuchsia. D ORFs identified in phage Sem4 are shown in dark blue

Comparative genomic analyses of the phages, including intergenomic relationships, phylogenetic positioning, and gene-level synteny. A Intergenomic similarity and distance heat map based on VIRIDIC analysis. Alignment genome fraction is shown as a gradient from salmon (low) to white (high); genome length ratio from black (low) to white (high); intergenomic similarity from white (low) to dark green (high); and intergenomic distance from dark blue (high) to white (low). B Phylogenomic tree of the four phages and closely related phages identified via BLASTn (≥70% coverage, ≥90% identity). C Circul

Full text

Introduction

Escherichia coli, while constituting a commensal inhabitant of the intestinal microbiota, encompasses pathogenic variants that have acquired an array of virulence determinants, enabling them to cause serious infections and widespread outbreaks, most frequently associated with microbial contamination of food and water (Hunter, 2003; Kunz et al., 2024; Centers for Disease Control and Prevention [CDC], 2024). Beyond its established role as an enteric pathogen, pathogenic E. coli has emerged as a critical indicator organism for fecal contamination in water and food safety assessments. Its frequent isolation from inadequately treated drinking water and food products not only underscores lapses in sanitation and hygiene but also exemplifies the challenges in mitigating waterborne and foodborne disease transmission (World Health Organization [WHO], 2018). Furthermore, its ability to survive and even thrive in diverse environmental matrices, ranging from agricultural runoff to biofilms in water distribution systems, amplifies its epidemiological relevance particularly in low-resource settings where infrastructure limitations exacerbate public health vulnerabilities (Forstinus et al., 2016; Leclerc et al., 2002; Meals & Braun, 2006; Muirhead, 2023). Pathogenic E. coli possesses the ability to induce a broad spectrum of infections, affecting multiple organ systems such as the gastrointestinal and urinary tracts, as well as the central nervous system, often leading to severe clinical manifestations depending on the pathogenic strain and host susceptibility (Bien et al., 2012; Gholizadeh et al., 2019).

The clinical impact of E. coli infections has been further complicated by the rise of antibiotic resistance, largely driven by the unregulated use of antimicrobials in human and veterinary medicine, compounded by poor hygiene practices and inadequate infection control measures (Pormohammad et al., 2019). Prolonged antibiotic exposure has fostered the emergence of multidrug-resistant E. coli strains, posing a significant public health threat. The development of extended-spectrum β-lactamases (ESBLs) has made E. coli resistant to a broad range of antibiotics, including penicillins, cephalosporins, and the monobactam aztreonam. Additionally, resistance to carbapenems, considered a last-resort antibiotic class, is also emerging, making treatment options even more limited (CDC, 2024; Paul et al., 2022). The proliferation of multidrug-resistant E. coli, particularly those producing ESBLs and carbapenemases, has raised significant public health concerns, as these strains contribute to life-threatening infections (Sáenz et al., 2004; Silva et al., 2020). These resistant strains are not only found in humans but are prevalent in animals and food products, facilitating transmission through direct contact or consuming contaminated food (Pormohammad et al., 2019; Samtiya et al., 2022). Environmental reservoirs, particularly hospital waste, contribute to the dissemination of antibiotic resistance genes, which are transmitted through sewage systems and into the broader environment, amplifying the spread of resistance (Pormohammad et al., 2019). The relentless escalation of resistance in E. coli, compounded by its ability to adapt to each new generation of antibiotics rapidly, underscores the urgent need for novel antimicrobials and innovative therapeutic strategies (Sabtu et al., 2015).

The employment of naturally occurring lytic bacteriophages for decontaminating antibiotic-resistant and susceptible pathogenic E. coli strains has emerged as a promising therapeutic and biocontrol strategy across various industrial sectors (Barros et al., 2019; Beheshti Maal et al., 2015; Brüssow, 2005; Huang et al., 2022; Oliveira et al., 2009; Wang et al., 2017). Phages exhibit strong antibacterial activity due to their high host specificity, rapid bacteriolysis, self-replication, and extracellular metabolic inertness (Dokuz et al., 2025; Melo et al., 2017; Oliveira et al., 2009; Ozbek et al., 2025), which makes E. coli phages particularly versatile biocontrol agents, applicable in food and water decontamination (Ahiwale et al., 2012; O’Flynn et al., 2004; Pereira et al., 2017; Zhou et al., 2022), agricultural biopesticides (Brüssow, 2005; Svircev et al., 2018), and veterinary and clinical therapies (Abdulamir et al., 2014; Bruttin & Brüssow, 2005; Dini & De Urraza, 2010; Zalewska-Piątek & Piątek, 2020). Among these, members of the Vectrevirus genus within the Autographiviridae family, exhibiting the podovirus morphotype, are known for their high specificity and their ability to selectively infect pathogenic E. coli strains (Asgharzadeh Kangachar et al., 2024; Markusková et al., 2024).

This study introduces the isolation and in-depth characterization of four previously unreported phages, namely E. coli phage Sem1 (Sem1), E. coli phage Sem2 (Sem2), E. coli phage Sem3 (Sem3), and E. coli phage Sem4 (Sem4). These phages were morphologically and genomically characterized, and their potent lytic activity was demonstrated in vitro against a multidrug-resistant clinical isolate Escherichia coli (MDR-Ec) strain. Furthermore, Sem4 exhibited remarkable biocontrol efficacy in simulated drinking water conditions, highlighting its potential as a natural and sustainable antimicrobial tool for waterborne pathogen management.

Materials & Methods

Bacterial Strains and Culture Collections

In this study, 18 bacterial strains from the Microbiology Laboratory collection of the Department of Molecular Biology and Genetics at Yıldız Technical University were used. These strains served to evaluate the lytic spectra of the isolated phages. Of these, MDR-Ec isolate and Klebsiella aerogenes ATCC 13,048 were used for both the isolation and amplification of phages. Originally used for Klebsiella phage screening, K. aerogenes was included due to its role in isolating the E. coli-infecting phage Sem4. Table 1 illustrates a detailed list of all strains used. Bacterial cultures were maintained at 37 °C in tryptic soy broth (TSB; Merck, USA), Mueller-Hinton broth (MHB; Merck, USA), on tryptic soy agar (TSA; Merck, USA), TSA soft agar overlays (0.6% agar), or Mueller-Hinton agar (MHA; Merck, USA), depending on the experimental requirements.

Table 1. Phages’ lytic spectra against tested bacterial strainsSpeciesStrain CharacteristicsInfectivitySem1Sem2Sem3Sem4 Escherichia coli Clinical-originated multidrug-resistant (MDR-Ec) (see Table S1)HostHostHost+EOP: HighATCC 10,536−−−−Adherent-invasive E. coli (AIEC)LF82, Clinical isolate−−−−Enterohemorrhagic E. coli (EHEC)Clinical isolate−−−−ATCC 35,218−−−−O157:H7−−−−302isolated from sea of Marmara−−−−P2/aisolated from water-cooling tower−−−−4/4isolated from Turkish raw meatball−−−− Klebsiella aerogenes ATCC 13,048+EOP: High+EOP: High+EOP: HighHostKa-1isolated from urine−−−− Acinetobacter baumannii ATCC B-AA-747−−−− Klebsiella pneumaniae ATCC B-AA-1144−−−− Pseudomonas aeruginosa ATCC 27853−−−−Salmonella Typhimurium−−−− Bacillus subtilis ATCC 6633−−−− Enterococcus faecalis ATCC 29212−−−− Streptococcus pyogenes Clinical isolate−−−− Staphylococcus aureus ATCC 25923−−−− S. aureus Methicillin-resistant, clinical isolate−−−−Ligolactobacillus salivarius (Bedir et al., 2025)C 1–4 isolated from chicken gastrointestinal system−−−−High EOP is defined as >10%

Antibiotic Resistance Profile

The antibiotic susceptibility profile of the clinical isolate (MDR-Ec) was determined using the Vitek-2 automated system with the AST-N420 card (bioMérieux, France). Resulted MIC patterns were interpreted based on the latest clinical breakpoints established by the European Committee on Antimicrobial Susceptibility Testing (EUCAST Version 15.0, 2025) (Table S1).

Isolation of Lytic Phages

For the isolation of E. coli phages, samples were collected from three distinct sources: wastewater from the Istanbul Water and Sewerage Administration (source of phage Sem1), wastewater from the Adana Water and Sewerage Administration (source of phage Sem2), and water trough samples from Karataş, Adana (sources of phages Sem3 and Sem4). To enrich for phages, 10 mL of each water sample was added to 90 mL of buffered peptone water (BPW; HiMedia, India), and incubated overnight at 37 °C with shaking at 180 rpm. The following day, samples were centrifuged at 5000×g for 10 min at room temperature, and the supernatants were filtered through sterile 0.22 μm pore-size CA membranes (Var et al., 2018). Subsequently, 10 mL of the filtrate was mixed with 100 µL of overnight cultures of either MDR-Ec (for phages Sem1, Sem2, and Sem3) or K. aerogenes ATCC 13,048 (for phage Sem4) grown in TSB, and incubated at 37 °C with shaking at 180 rpm. After incubation, samples were centrifuged at 10,000×g for 10 min at 4 °C, and the supernatants were again filtered through sterile 0.22 μm CA membranes. At this stage, the resulting filtrates were expected to be highly enriched in lytic phages active against E. coli. Spot assays were performed to assess phage-host specificity in the obtained filtrates. The filtrates were spotted onto TSA plates overlaid with overnight cultures of the host bacteria and incubated at 37 °C until clear zones (indicative of bacterial death) appeared. Subsequently, four distinct filtrates exhibiting inhibition zones were serially diluted and subjected to drop assays, both to confirm the presence of phages through plaque formation and to enable plaque isolation. Individual plaques were carefully excised using a sterile scalpel and subjected to five rounds of enrichment in TSB and re-plating, in order to ensure the isolation of phages with uniform single-plaque morphology (Erdogdu & Ozbek, 2025).

Phage Concentration and Purification

Phage particles were propagated using the double-layer plaque assay with minor modifications (Tasdurmazli et al., 2023). Briefly, 100 µL of phage filtrate and 100 µL of host bacterial culture were mixed with 5 mL of 0.6% TSA-soft agar and overlaid onto TSA plates. Following complete bacterial lysis, 5 mL of SM buffer [100 mM NaCl, 8 mM MgSO₄, 50 mM Tris-HCl (pH 7.5), 0.002% (w/v) gelatin] was added to each plate. The plates were incubated with gentle agitation at 60 rpm on a 3D Sunflower Mini-Shaker (BIOSAN) at 4 °C for 18 h to facilitate the transfer of phage particles into the buffer (Melo et al., 2014). The resulting mixture of liquid and soft agar was collected, centrifuged at 10,000×g for 10 min at 4 °C, and the supernatant was filtered through a 0.22 μm membrane as previously described. The phage lysate was precipitated using NaCl (5.84% w/v) and polyethylene glycol 8000 (10% w/v), followed by purification with chloroform (1:4, v/v), and stored at 4 °C until further use (Melo et al., 2018).

Morphological Characterization

Purified phage suspensions containing ~ 10⁸ PFU were carefully deposited onto carbon-coated Formvar films mounted on 200-mesh nickel grids. Following adsorption, grids were negatively stained with 2% (w/v) uranyl acetate (pH 4.0) and examined under a JEOL JEM-1400 transmission electron microscope (JEOL Ltd., Tokyo, Japan) operating at standard accelerating voltage conditions (Melo et al., 2018).

Host Range Profiling and Efficiency of Plating

Phage lytic activity was assessed by spot assay as described above. Phage suspensions (10⁸ PFU/mL) were applied onto TSA plates seeded with the bacterial strains detailed in Table 1. For strains exhibiting visible lysis, 10-fold serial dilutions of the suspensions were subsequently spotted on the bacterial lawns. After 24 h incubation at 37 °C, lytic profiles were recorded (Selcuk et al., 2024). Relative efficiency of plating (EOP) was scored as the ratio of the phage titre obtained on each target strain to that on the original propagation host (Barros et al., 2019).

Kinetic Profiling of Phage Infectivity Through One-Step Growth Analysis

One-step growth curve (OSGC) experiments were conducted with minor modifications to previously described protocols (Melo et al., 2014). In brief, 5 mL of MDR-Ec culture at mid-exponential phase (OD₆₀₀ ≈ 0.5) was harvested by centrifugation (5 min, 7000×g, 4 °C), resuspended in 5 mL of fresh TSB, and subsequently infected with 5 mL of phage suspension at a multiplicity of infection (MOI) of 0.01. Phages were allowed to adsorb onto the bacterial cells for 5 min at 37 °C with gentle agitation (120 rpm). The suspension was then centrifuged under the same conditions, and the resulting pellet was resuspended in 10 mL of fresh TSB. Samples were collected at 5-minute intervals for the first 30 min post-infection, and subsequently every 10 min until 70 min, for further analysis. All experiments were independently conducted in triplicate. Latent period and burst size were calculated as indicators of phage replication dynamics, with all assays performed in independent biological triplicates to validate consistency.

Temporal Profiling of Phage Dose-Dependent Antibacterial Activity Through Bacterial Turbidity Measurements

To evaluate the impact of phage dose on bacterial growth, MDR-Ec was cultured to the exponential phase in TSB. The bacterial culture was mixed with phage suspensions at MOIs of 0.01, 0.1, 1.0, and 10, and dispensed into 96-well microplates in triplicate. Bacterial growth was continuously monitored by measuring optical density at 600 nm (OD₆₀₀) at 30-minute intervals over an 18-hour incubation period using a microplate spectrophotometer (Epoch 2 BioTek, USA). All assays were independently repeated three times. Phage-free bacterial cultures and cell-free phage suspensions were included as controls to validate the assay (Danis-Wlodarczyk et al., 2020).

Impact of Environmental Conditions on the Infective Potential and Stability of Phage Filtrates

To evaluate the pH stability of the phage filtrates, they were incubated with a universal pH buffer (150 mM potassium chloride, 10 mM potassium dihydrogen phosphate, 10 mM sodium citrate, 10 mM boric acid) at pH values of 2, 4, 6, 8, 10, and 12, maintaining a final concentration of 5 × 10⁸ PFU/mL at a 1:10 (v/v) ratio. The mixtures were incubated at 25 °C for 2 h. After incubation, serial dilutions of the samples were performed and plated to assess phage viability. For temperature stability testing, phage filtrates, initially concentrated at 5 × 10⁸ PFU/mL, were subjected to incubation at temperatures of 4, 25, 37, 45, 65, and 90 °C for 2 h. Following incubation, dilutions were prepared, and phage activity was assessed. To evaluate the long-term stability of phage suspensions at a concentration of 5 × 10¹¹ PFU/mL in SM buffer, samples were stored at refrigeration temperature. Phage activity was monitored at regular intervals over a 14-month period. All assays were performed independently in triplicate.

Evaluation of Phage-Mediated Inhibition of MDR-Ec Growth Through Minimum Inhibitory Concentration Testing

The antimicrobial efficacy of the phages against MDR-Ec was assessed by determining the minimum inhibitory concentration (MIC) through the broth microdilution method. To investigate the dynamics of phage-bacterium interactions in different media, this study was conducted using two distinct culture media. Briefly, the bacterial culture, separately cultivated in MHB and TSB, was adjusted to 10⁶ CFU/mL and co-incubated with phage suspensions and phage cocktail (for 100 µL sample volume, all four phages were prepared in equal concentrations and amounts) prepared at varying concentrations (final concentrations of 5 × 10^1^, 5 × 10³, 5 × 10⁵, and 5 × 10⁷ PFU/mL in each well) in a 96-well microplate. Bacterial growth inhibition was assessed by spectrophotometric analysis following an 18-hour incubation at 37 °C without agitation. Three experiments were conducted in triplicate to ensure reproducibility (Erdogdu & Ozbek, 2025).

Phage-Mediated Decontamination of Tap Water Contaminated with MDR-Ec

In this experiment, the work of Kauppinen et al. (2021) was utilized with some modifications Metin girmek için buraya tıklayın veya dokunun. The test water used in this study was the municipal tap water from Esenler, Istanbul. This water is sourced from surface water resources and undergoes disinfection through ozonation, with chlorine being applied before distribution (for physical-chemical properties of the water, refer to Table S2). To determine the efficacy of phages in eliminating MDR-Ec from tap water, 50 mL of filtered tap water was inoculated with the bacteria at a final concentration of approximately 10^7^ CFU/mL, and phage Sem4 was added at an MOI of 1. The suspension was incubated in an Erlenmeyer flask at room temperature, and bacterial decontamination was monitored at multiple time points (0, 4, 8, 24, and 48 h). At each time point, the absorbance of the solution was measured, and colony counts were determined by plating on MHA. In the control experiment, SM buffer was added instead of phage. The experiments were repeated in triplicate to ensure statistical reliability and consistency of the results.

Phage Genome Sequencing, Prediction of Coding Sequences, and Functional Annotation

The phages DNA was extracted using the Phage DNA Isolation Kit (Norgen) following the instructions provided by manufacturer, and the concentration and purity of the DNA eluents were determined by the nanodrop spectrophotometer (NanoDrop One Microvolume UV-Vis Spectrophotometer, Thermo Fisher Scientific) and Qubit 2.0 (Thermo) with Qubit dsDNA HS Assay Kit (Invitrogen, Cat. No. Q32854). Whole-genome sequencing of phages’ DNA and downstream NGS processing were performed according to the protocol established by Erdogdu and Ozbek with modifications such as annotation tools comparison and manual ORF checks (Erdogdu & Ozbek, 2025). Annotation of the phage contigs were carried out using PHANOTATE via BV-BRC Genome Annotation Service (BV-BRC, www.bv-brc.org) (Mcnair et al., 2019), and Pharokka version 1.3.2 via Galaxy Phage with default parameters (Bouras et al., 2023; Ramsey et al., 2020). The results of the annotation were manually reviewed using sequence comparisons against the NCBI protein database. Pseudo-ORFs identified by phage annotation tools were corrected manually, and only the curated coding sequences—excluding pseudo-ORFs—were submitted to NCBI. The presence of tRNA genes was evaluated by both annotation tools through tRNAscan-SE. Additionally, potential proteins with depolymerase activity were predicted using the Galaxy phageDpo tool (Galaxy | Galaxy Docker Build). Proksee server was used to create a linear map of the isolated phage genomes. Antibiotic resistance, toxins, and other virulence genes were investigated with the Proksee Server CARD (Comprehensive Antibiotic Resistance Database) tool and VRprofile2 web server (Alcock et al., 2020; Wang et al., 2022).

Phages’ Phylogenetic Relationship and Proteomic Comparisons

The phage genomes analysis and phylogenetic relationship were conducted by using the different bioinformatics tools according to Erdogdu et al. (2025). NCBI BLASTn was used to identify phages with high sequence similarity. Whole-genome-based phylogenetic analysis of the phages was performed using the VICTOR tool (Meier-Kolthoff & Göker, 2017), based on sequences retrieved from the NCBI nucleotide database with a minimum query coverage of 70% and sequence identity of at least 90% (NCBI Nucleotide, accessed 20 June 2025). All pairwise comparisons of nucleotide sequences were conducted using the Genome BLAST Distance Phylogeny (GBDP) method, applying the settings recommended for prokaryotic viruses. The resulting intergenomic distances were used to construct a balanced minimum evolution tree with branch support using FASTME, including SPR post-processing with the D_0_ formula.

Comparative genomic similarity between the isolated phages and related phages was identified through NCBI blastn and VICTOR analysis, assessed using Viridic Web (Moraru et al., 2020), and the results were visualized as a heat map. Assembly data for the phages were retrieved from the NCBI Datasets (Datasets - NCBI - NLM), and the heat maps were generated using the ggplot2 package. Default VIRIDIC thresholds of 95% and 70% nucleotide similarity were used to define species- and genus-level relationships, respectively.

Proteomic comparisons of phages were conducted between the isolated phages and one of the most closely related phages of the same species using the BV-BRC Proteome Comparison service with both unidirectional and bidirectional alignment approaches. The analysis was performed using default parameters, which rely on bidirectional best hits (BBHs) identified through BLASTP comparisons, which align all protein-coding sequences (CDSs) between the reference and query genomes to determine functional similarity and conservation (BV-BRC, accessed 12 July 2025).

Nucleotide Sequence Accession Numbers

The complete genome sequences of the phages have been deposited in GenBank under the accession numbers PV775411, PV775410, PV775413, and PV717788.

Statistical Analysis

Data analysis was performed using IBM SPSS Statistics version 30.0.0.0 (IBM Corp., Armonk, NY, USA). One-way ANOVA followed by Tukey’s post hoc test was used to compare independent groups when more than two groups were involved. An independent samples t-test was conducted to assess differences between two independent groups (e.g., control and treatment), assuming normal distribution of the data. Statistical significance was defined as < 0.05. Graphs were generated using Origin 2021 (OriginLab Corp., Northampton, MA, USA).

Results

Antibiotic Susceptibility Testing Revealed That the Clinical E. coli Isolate is Multidrug-Resistant

The strain exhibited resistance to a broad range of β-lactam antibiotics, including penicillins (ampicillin, amoxicillin/clavulanate, piperacillin/tazobactam), cephalosporins from the 2ⁿᵈ to 4ᵗʰ generation (cefuroxime, cefotaxime, ceftazidime, ceftriaxone, cefepime), and the carbapenem ertapenem (Table S1). In contrast, susceptibility was retained to imipenem, meropenem, amikacin, tigecycline, and trimethoprim/sulfamethoxazole. Notably, ciprofloxacin was classified as resistant under meningitis-specific EUCAST breakpoints, but intermediate under other conditions (EUCAST Version 15.0, 2025). These results suggest extensive resistance, with therapeutic options limited to a few last-resort antibiotics.

E. coli Phages Show Specific Lytic Spectra, with Sem4 Displaying Enhanced Replicative Capacity

Phages infecting E. coli are commonly found in environments contaminated with fecal matter, such as wastewater and environmental waters (WHO, 2018; Forstinus et al., 2016); therefore, these sources were selected for phage isolation. After five rounds of purification, four phages were isolated from three different sources. Following five rounds of purification, four phages were isolated from three distinct sources. These three phages formed plaques with an average diameter of 4.0 ± 0.1 mm, characterized by wide, opalescent halos and defined lysis centers (Fig. 1A-i—iii), and exhibited relatively similar plaque morphologies. In contrast, the plaque morphology of the fourth phage (Fig. 1A-iv) on its host bacterium also featured opalescent halos and defined lysis centers, but had a smaller average diameter of 2.7 ± 0.3 mm. Although initial screenings were performed using 0.5% agar concentration, the medium was subsequently adjusted to 0.6% to better accommodate the large plaque diameters. Phages were named based on their source and order of isolation (see Methods 2.2) and are hereafter referred to as Sem1, Sem2, Sem3, and Sem4 throughout the manuscript. Electron micrographs of negatively stained preparations revealed that phages Sem1, Sem2, Sem3, and Sem4 displayed typical podovirus morphology, with short, non-contractile tails (Fig. 2). Each phage possesses an icosahedric capsid with 56 nm in diameter, and a tail measuring 17 nm in length and 27 nm in width.

Fig. 1. In vitro characterization of phages Sem1, Sem2, Sem3, and Sem4 (i–iv, respectively). A Representative plaque morphology of the phages formed on bacterial lawns of their respective hosts. B Transmission electron micrographs of negatively stained phage particles (2% [w/v] uranyl acetate), displaying typical podovirus morphology; scale bars represent 0.1 μm. C OSGCs of the phages on MDR-Ec. Y-axis denotes log₁₀-transformed phage titers at distinct time points. Error bars represent standard deviations from three independent replicates (= 3). D The effect of different MOIs of the phages on the growth of MDR-Ec, where MOI 10 (■, black), MOI 1 (●, wine), MOI 0.1 (▲, dark cyan), MOI 0.01 (▼, purple), and the uninfected control (♦, grey). Although OD₆₀₀ measurements were recorded every 30 min during the 18-hour incubation at 37 °C, data are presented at 2-hour intervals for clarity. The blank control exhibited an absorbance value of approximately 0.08–0.09. Statistically significant reductions in OD compared to the untreated control are indicated as < 0.001 (∗∗∗). Error bars represent the standard deviation (= 3)

Fig. 2. Stability and antibacterial activity assays of phages Sem1, Sem2, Sem3, and Sem4, and biocontrol performance of Sem4 in tap water. A–C Stability profiles of the phages under different pH conditions (A), temperatures (B), and long-term storage (C; percentages indicate the relative change in phage titers at the end of the incubation period) (= 3). D Minimum inhibitory concentration assay results for the four phages across two growth media (TSB and MHB) and four phage concentrations (top to bottom: 5 × 10^7^, 5 × 10^5^, 5 × 10^3^, and 5 × 10^1^ PFU/mL), measured via OD_600_ absorbance at the end of incubation. Each bar represents the average of three technical replicates; medium controls exhibited OD_600_ values of approximately 0.08–0.09. E Biocontrol efficiency of Sem4 in tap water over 48 h, showing changes in absorbance (line graph: purple circles for uninfected control, purple stars for Sem4-treated) and bacterial counts (bar graph: grey for uninfected control, white for Sem4-treated). The blank control consistently showed OD_600_ values around 0.08–0.09. Statistically significant reductions in absorbance compared to the untreated control are indicated as <0.05(∗) and <0.001(∗∗∗). Error bars represent standard deviations (= 3). (F) Visual comparison of turbidity in Erlenmeyer flasks after 48 h, showing uninfected (control) and Sem4-treated tap water samples

The lytic spectra of the phages against 21 bacterial strains are presented in Table 1. All phages exhibited equally high EOP against MDR-Ec, the host of Sem1, Sem2, and Sem3, as well as K. aerogenes ATCC 13,048, the host of Sem4. Given this observation, considering the critical clinical relevance of the MDR-Ec bacterium, this strain was selected as the primary target organism for all subsequent analyses. Overall, when assessing the host lysis performance of each phage across the remaining 19 strains, including pathogenic E. coli isolates (Table 1), none of the phages demonstrated lytic activity. This indicates that the phages infecting Gram-negative bacteria possess a narrow host range. However, it should be noted that the host range observed in laboratory assays may not fully reflect the theoretical host spectrum, as it was constrained by the limited panel of tested strains. Expanding future screening efforts to include a broader diversity of E. coli isolates from culture collections could substantially enhance the comprehensiveness and ecological relevance of host range assessments. Furthermore, evaluations based on genomic similarity were constrained by recent taxonomic revisions by the ICTV, which have reassigned previously described Vectrevirus phages to different genera, thereby complicating comparative analyses (Ferriol-González et al., 2024; Ferriol-González & Domingo-Calap, 2024; Mizuno et al., 2020; Turner et al., 2025).

One-step growth curve analysis, a standard approach in phage biology, was performed to investigate the life cycle and replication kinetics of the phages. Using MDR-Ec as the host strain, both the latent period and burst size were determined. Detailed examination of the growth curves presented in Fig. 1C revealed that phage Sem1 exhibited a latent period of 15 min and released an average of 84 phage particles per infected cell following a single infection cycle (Fig. 1C-i). For phage Sem2, the latent period was 20 min with a burst size of 83 particles per cell (Fig. 1C-ii), while Sem3 displayed a similar profile with a 20-minute latent period and a burst size of 80 particles per cell (Fig. 1C-iii). In contrast, Sem4 exhibited a comparable latent period of 20 min but a significantly higher burst size of 429 particles per cell (Fig. 1C-iv).

Interactions between MDR-Ec and phages were evaluated at four distinct levels of MOI. Spectrometric fluctuations in OD values at 600 nm, resulting from the density of live bacterial cells, were recorded over 18-hour in relation to phage infection (Table S3) and graphed accordingly (Fig. 1D). As shown in Fig. 1D-i, Sem1 inhibited MDR-Ec growth in a concentration-dependent manner, with the fastest decline at MOI 10 within the first 10 h (from ~ 0.5 to 0.14), followed by MOI 1, while MOIs 0.1 and 0.01 produced similar reductions (> 0.05). Sem2 (Fig. 1D-ii) showed an initial decrease at MOI 10, but OD₆₀₀ rose to ~ 0.4 by 18 h, and at MOIs 0.1 and 0.01 its lytic activity exceeded that of Sem1 (< 0.05). For Sem3 (Fig. 1D-iii), OD₆₀₀ remained lowest at MOI 10 by the 7th hour, and lytic activity at MOIs 0.1 and 0.01 was similar (> 0.05). Sem4 (Fig. 1D-iv) showed MOI 1 kinetics comparable to MOI 10 (> 0.05), with OD₆₀₀ decreasing to 0.17 by 12 h at MOI 10 and increasing slightly to 0.20 thereafter. At MOI 0.1, Sem4 produced the strongest reduction among the phages (~ 0.27 by 12 h), while at MOI 0.01 an early, sharper decline occurred despite similar final values. Across all phages, OD₆₀₀ generally increased during the first two hours at MOIs 0.1 and 0.01, except for Sem4 at MOI 0.1 and for the modest increase observed at MOI 10 during incubation. Overall, Sem4’s profiles were consistent with OSGC results, confirming its superior infectivity.

E. coli Phages Remain Infective Across Diverse pH and Temperature Conditions, with Sem4 Efficiently Eliminating MDR-Ec from Drinking Water

Temperature and pH stability are critical parameters for determining the infective range of phages under normal or extreme conditions, especially when considering their potential therapeutic or biocontrol applications. As shown in Fig. 2A, the phage titers were determined against MDR-Ec following a 2-hour incubation at various pH levels. No significant change in infective activity was observed for any of the four phages at pH 8 and 10. However, Sem4 exhibited a relatively reduced activity at pH 10 compared to pH 8, and this reduction was more notable when compared to the other phages. For all phages except Sem3, infectivity at pH 4 and pH 6 decreased relative to pH 8. Interestingly, Sem3 displayed an unusual trend: its activity increased as the environment became more acidic, suggesting enhanced stability or infectivity at lower pH values. No infective activity was detected for any of the phages at the extreme pH values of 2 and 12. Overall, aside from the extreme conditions, the phages remained stable under both mildly acidic and basic environments, with Sem3 being particularly notable for its increased activity at pH 4 and 6. Aside from this, the phages exhibited broadly similar stability profiles. In Fig. 2B, the infectivity of phages was assessed after pre-incubation at various temperatures. Within the range of 4 °C to 45 °C, all phages retained nearly full infective capacity. At 65 °C, although Sem3 and Sem4 appeared slightly more stable, all phages exhibited an average activity loss of approximately 40–45%. At 90 °C, none of the phages maintained infectivity, clearly indicating thermal inactivation at this extreme temperature. The long-term stability of the phages was assessed by storing them at 4 °C in SM buffer with an initial titer of 5 × 10^11^ PFU/mL. Over 14 months, phage titers exhibited only minor reductions, 6.09% for Sem1, 8.7% for both Sem2 and Sem3, and 11.3% for Sem4. Titers remained largely unchanged during the first 6 months of storage (data not shown), and overall results indicate that all four phages retained substantial stability under standard refrigeration conditions, supporting their applicability as durable biocontrol agents in industrial and therapeutic contexts.

Minimum inhibitory concentration analysis was performed after 18 h of incubation, with the average absorbance values of the wells at 600 nm (n = 3) recorded. Different antibacterial activity levels of the phages and the quadruple phage cocktail at four distinct concentrations were determined against MDR-Ec (Fig. 2D). The reading of the bacteria wells without phage, for TSB and MHB media, showed average OD_600_ values of 0.929 and 0.820, respectively, with readings for both MHB and TSB media at 600 nm ranging from 0.08 to 0.09. When results were examined sequentially, for Sem1, growth was observed only at 5 × 10^1^ and 5 × 10^3^ PFU/mL, with inhibition effects at 5 × 10^3^ PFU/mL being significant (OD_600_ values of 0.283 for TSB and 0.271 for MHB), while at 5 × 10^1^ PFU/mL, growth remained high with OD_600_ values of 0.782 and 0.695, respectively. For Sem2, growth was only observed at the lowest phage concentration in MHB, where an OD_600_ of 0.547 was recorded, while in TSB, growth was present at the two lowest phage concentrations, with 5 × 10^3^ PFU/mL showing growth at a very low OD600 of 0.159. For Sem3, antibacterial activity was higher compared to the two previously discussed phages; in MHB, the lowest phage concentration resulted in a remarkably low absorbance value of 0.136, while in TSB, growth was observed only at the lowest dose with relatively low antibacterial effect (OD_600_ = 0.633). Sem4 demonstrated the best results among the four phages, which corroborated the findings from the initial in vitro analyses (Fig. 1C-iv, D-iv). No growth was observed in MHB under any circumstances, while in TSB, the lowest phage concentration resulted in an OD_600_ reading of 0.394, the lowest observed among the other phages. Upon reviewing all the results, no growth was detected at 5 × 10^5^ and 5 × 10^7^ PFU/mL for any of the phages in any medium. A comparison between the same concentrations in TSB and MHB revealed consistently better activity in MHB. Among all the phages tested, Sem4 exhibited the highest inhibitory activity, with an MIC value below 5 × 10^1^ PFU/mL. For the phage cocktail, an antagonistic effect was observed at the lowest phage concentration, as readings in the same condition were higher than those for Sem4, resulting in a lower antibacterial effect.

Since Sem4 exhibited the highest activity in all previous antibacterial assays, it was selected for biocontrol of MDR-Ec in tap water. As shown in Fig. 1E, OD_600_ measurements revealed that in the control group, absorbance increased markedly from 0.091 to 0.15 within 24 h and slightly decreased to 0.135 by 48 h. In contrast, in the phage-treated group, absorbance rose modestly from 0.091 to 0.115 within the first 4 hours, followed by a sharp decline to 0.090 by 24 h, which was maintained through 48 h. This reduction was also visually evident through a marked decrease in turbidity, which was photographed and presented in Fig. 1F. Colony counts (Fig. 1E) further supported these findings: while the control group increased from an initial 10^7^ CFU/mL to 3 × 10^8^ CFU/mL, the phage-treated group dropped to 7 × 10^4^ CFU/mL by 8 hours, and no colonies were detected at 24 and 48 h. These results demonstrate that Sem4 is a highly effective candidate for the decontamination of MDR-Ec in tap water.

Genomic Architecture and Predicted Functional Modules of the Isolated Phages

Nanopore reads were de novo assembled using Flye (v2.8) with a minimum 30× coverage, producing complete genomes as single contigs. The E. coli phage genomes ranged from 44,244 to 45,205 bp, contained 66–74 coding sequences, exhibited 44–45% GC content, and lacked tRNA genes (Table 2).

Table 2. Genome features of isolated phagesAnnotation statisticalSem1Sem2Sem3Sem4Length (bp)44,71144,24445,13745,205Average Coverage120122517GC Content(%)44.9244.9745.0145No. of CDS66746873No. of tRNAs0000

When we compared the predicted number of CDSs with those of the most closely related phages available in the NCBI RefSeq database (see SuppFile1–4, 6), we observed that the tools Pharokka and PHANOTATE tended to generate a higher number of redundant CDS predictions. We manually excluded these CDSs if: (i) their BLASTp results against the NCBI database showed insufficient query coverage or sequence similarity, or (ii) the CDS search returned “no significant similarity”. Additionally, if a CDS was split or contained non-canonical start or stop codons, it was flagged as a misprediction. The final dataset was curated to retain only validated CDSs; locus tags were manually assigned, annotation features (e.g., gene product descriptions) were revised (Grigson et al. 2023; Turner et al. 2021a, b), and the curated genomes were subsequently submitted to NCBI under the following accession numbers: PV717788, PV775410, PV775411, and PV775413. Linear genome organization was visualized using Proksee tools based on the GenBank files of the manually curated genomes. The identified functional genes include those related to structural components (tail, head, and packaging), host cell lysis, and DNA/RNA/nucleotide metabolism, as well as phage defense (SuppFile11). In the linear genome visualizations generated via Proksee, functionally annotated proteins are labeled, whereas hypothetical proteins remain unlabeled (Fig. 3). The number of identified proteins per phage was 24 (Sem1), 29 (Sem2), 23 (Sem3), and 30 (Sem4), respectively (Table S4). According to CARD analysis, no antimicrobial resistance genes were detected in any of the phage genomes (Fig. 3).

Fig. 3. Linear genomic representation of the phages. CARD analysis results are indicated in red. GC content, GC skew, and ORFs are shown in distinct colors. A ORFs identified in phage Sem1 are shown in fuchsia. B ORFs identified in phage Sem2 are shown in purple. C ORFs identified in phage Sem3 are shown in fuchsia. D ORFs identified in phage Sem4 are shown in dark blue

Phylogenomic Characterization of Vectrevirus-Related E. coli Phages: Conserved Taxonomy, Divergent Genomic Organization

To determine the taxonomic relationships of the phages both among themselves and with previously characterized phages, multiple platforms were employed. Genomic similarity and inter-phage distances were assessed using VIRIDIC with default parameters. Based on the ICTV taxonomic classification criteria, all four phages were classified as belonging to the same species at the nucleotide level (Fig. 4). The NCBI BLASTn tool was used to identify related phages, and the results are summarized in SuppFile5. Phylogenetic trees were generated using the VICTOR platform, incorporating BLASTn data and applying consistent thresholds across all four phages (≥ 70% query coverage and ≥ 90% identity).

Fig. 4. Comparative genomic analyses of the phages, including intergenomic relationships, phylogenetic positioning, and gene-level synteny. A Intergenomic similarity and distance heat map based on VIRIDIC analysis. Alignment genome fraction is shown as a gradient from salmon (low) to white (high); genome length ratio from black (low) to white (high); intergenomic similarity from white (low) to dark green (high); and intergenomic distance from dark blue (high) to white (low). B Phylogenomic tree of the four phages and closely related phages identified via BLASTn (≥70% coverage, ≥90% identity). C Circular comparative map of protein-coding genes between the reference phage and closely related isolated phages (Sem1, Sem2, Sem3, and Sem4)

Figure 4B shows the phylogenomic GBDP tree inferred using the formula D0, and yielding an average support of 1%. The numbers above branches are GBDP pseudo-bootstrap support values from 100 replications. The branch lengths of the resulting VICTOR trees are scaled in terms of the respective distance formula used. OPTSIL clustering based on VICTOR analysis identified 35 species-level clusters (D0), while only one cluster was formed at both the genus and family levels (D0) (Fig. 4B). According to VICTOR, the phages represent a novel species within the Vectrevirus genus. To further support this, a VIRIDIC heat map was generated using Vectrevirus sequences available in NCBI. The results showed that the phages did not exceed 95% intergenomic similarity with any known Vectrevirus, confirming their classification as a new species according to ICTV thresholds (Supp_Material_Our phages and all NCBI submitted Vectrevirus).

Proteomic comparisons between the isolated phages and a related phage, Escherichia phage EEc4, were performed using the BV-BRC PATRIC platform, focusing primarily on proteins with defined functions. Among the annotated ORFs, the capsid assembly scaffolding protein Gp9 (peg41) of EEc4 showed over 90% sequence similarity to its counterparts in Sem1 (accession XUY71613.1, locus tag TOLAB_0064) and Sem3 (accession XWX23438.1, locus tag TOLAB3_0026). In contrast, the corresponding proteins in Sem2 (accession XWX23839.1, locus tag TOLAB2_0046) and Sem4 (accession XWX23509.1, locus tag TOLAB4_0030) showed lower degrees of similarity, reflects the diversity of isolated phages. The major capsid protein Gp10A (peg42), on the other hand, demonstrated approximately 80% sequence similarity across all four isolated phages. For the acetyl-CoA acetyltransferase, the EEc4 homolog peg52 showed about 40% identity with Sem2 (XWX23856.1, TOLAB2_0063) and approximately 60% identity with the other isolates. Another homolog, peg54, shared only ~ 30% identity with all compared phages (SuppFile11). Notably, tail spike proteins and acetyl-CoA acetyltransferases from the isolated phages aligned unidirectionally with the reference genome proteins peg61 and peg54, respectively (SuppFile11, Fig. 4C). Overall, protein-level comparisons between the isolated phages and the reference genome revealed consistent patterns of both sequence conservation and divergence (Fig. 4C).

Another important proteomic finding is related to depolymerase activity, where PHANOTATE predicted the proteins with IDs XWX23484 and XWX23485 in Sem4, XUY71565 and XUY71566 in Sem1, XWX23863 and XWX23864 in Sem2, and XWX23417 and XWX23418 in Sem3 as hypothetical proteins, whereas PHAROKKA annotated these sequences as tail fiber proteins (SuppFile11). On the other hand, PhageDPO predicted these sequences as putative depolymerases with high confidence, showing 96–99% probability scores (SuppFile7-10). Notably, PhageDPO analysis predicted these same sequences as putative depolymerases with high confidence, assigning probability scores ranging from 96% to 99% (SuppFile7-10). We identified two types of Dpo domains: one containing Pgu1 and β-helix motifs, and the other associated with the Peptidase S74 family. All isolates shared identical Dpo sequences among themselves, and these sequences showed high amino acid similarity to those from other Vectrevirus phages (e.g., 99.65% or 99.21% identity).

Discussion

The isolated phages target MDR Escherichia coli and K. aerogenes (ATCC 13048), both members of the phylum Pseudomonadota, a major lineage of Gram-negative bacteria. Bioinformatics analyses confirmed this taxonomy, and based on family-level classification, the phages were determined to infect members of the Enterobacteriaceae family (Markusková et al., 2024). Morphologically, these phages belong to the podovirus morphotype and the Autographiviridae family, which encompasses double-stranded DNA phages commonly isolated from diverse environments such as soil, water, and the human gut (Dalmasso et al., 2016; O’Flynn et al., 2004; Oliveira et al., 2009). This environmental ubiquity explains the ease of isolating such phages from wastewater and environmental water sources. Notably, all isolated phages formed clear plaques with surrounding halo zones, a morphology consistent with previously reported E. coli phages within the Vectrevirus genus (Markusková et al., 2024), suggesting the presence of a polysaccharide depolymerase domain within their receptor-binding proteins (RBPs). Host range analyses showed that the Escherichia phages confirmed by genomic data were only active against MDR-Ec strains (the host of Sem1, Sem2, and Sem3) and K. aerogenes ATCC 13,048 (the host of Sem4). The ability of Sem1, Sem2, Sem3, and Sem4 phages to infect both species, which are phylogenetically related, while lacking infectivity towards other tested strains such as AIEC, EHEC, and E. coli O157:H7, indicates a narrow host range. The observation that certain polyvalent phages, although active against multiple bacterial genera, nonetheless exhibit narrow host ranges is intriguing, though not entirely unexpected in light of existing literature (Chung et al., 2023). Host range analyses have shown that phage KFS-EC3, despite its ability to infect E. coli O157:H7, Salmonella spp., and Shigella sonnei, lysed only 7 out of 57 tested strains (Kim et al. 2021a, b). Similarly, phage S144 exhibited limited infectivity within Salmonella (9/72 strains), yet demonstrated broader polyvalence across a screened panel of 211 Enterobacteriaceae strains, including Cronobacter sakazakii and Enterobacter spp. (Gambino et al., 2020). Notably, even coliphages T4 and AR1, sharing a highly conserved core genome, have been reported to differ substantially in host range (Hamdi et al., 2017). A 2021 study reported phages capable of infecting both E. coli and Enterobacter cloacae, while failing to lyse certain E. coli strains (Addablah et al., 2021). Notably, such observations must be interpreted in light of the diversity and number of strains tested, which can significantly influence host range conclusions. The evolution of such polyvalent phages likely involves complex selective pressures, as all stages of infection must be functionally compatible with multiple hosts. In the context of our study, the presence of autonomous DNA polymerases in the phages may facilitate independent replication, whereas the absence of tRNA genes, combined with a marked codon usage compatibility with the host’s tRNA pool, suggests a potential translational adaptation to their bacterial hosts (Kim et al. 2021a, b). Although Enterobacteriaceae members likely share similar tRNA repertoires, inter-genus variations in codon usage and tRNA gene copy number may impact translational efficiency and phage adaptation (Rojas et al., 2018). Exploring shared translational biases across bacterial genera may thus shed light on host adaptation mechanisms. Polyvalent phages characterized in the literature typically broaden their host range by targeting conserved surface structures or by encoding multiple RBPs (Kim et al. 2021a, b; Markusková et al. 2024). Additionally, depolymerases warrant attention for their roles in host recognition and entry, which ultimately shape the breadth of the phage lytic spectrum (Knecht et al., 2020). The ensuing sections of this study provide a thorough examination of these aspects of the E. coli phages. Given the clinical importance of MDR-Ec strain, particularly due to its broad resistance profile, including penicillins, 2ⁿᵈ to 4ᵗʰ generation cephalosporins, and the carbapenem ertapenem, and its involvement in both community and hospital-acquired infections, it is unsurprising that subsequent experiments focused on these strains as host models.

Despite the genomic similarities reported for the four phages, notable differences in burst sizes were observed. However, considering the complex nature of phage biology and the high variability in phage-host and phage-environment interactions, such variations are not entirely unexpected. López-Cuevas et al. (2011) reported a similar scenario in which closely related E. coli phages isolated from cattle and poultry feces exhibited burst sizes ranging from 154 to 426 PFU/cell (López-Cuevas et al., 2011). A review of the literature indicates that latent periods of E. coli phages of the Autographiviridae family typically range from 15 to 20 min, which is consistent with our findings, and burst size values generally range from 50 to 200 PFU per infected cell (Guo et al., 2021; Markusková et al., 2024; Sun et al., 2022; Yuan et al., 2021). However, Sem4 exhibited a notably higher burst size of 429 particles per cell compared to previously reported values, which may account for its superior performance in antibacterial assays. Burst size may be affected by various factors, including phage genome features, receptor recognition efficiency, replication dynamics, endolysin gene expression, and the physiological state of the host cell (Clokie & Kropinski, 2009; Dennehy & Abedon, 2021; Hyman & Abedon, 2010; Kasman et al., 2002; Young, 1992). Any of these parameters may favor Sem4, presenting a promising subject for future investigation. Additionally, detailed protein comparison analyses (Fig. 4C and SuppFile11) also reflect the differences of SEM4 in both the reference genome and each other.

All tested phages effectively inhibited MDR-Ec proliferation, with suppression generally increasing at higher phage concentrations (< 0.001). Among them, Sem4 consistently demonstrated the strongest and most consistent bacteriolytic activity across a range of MOIs. This observation is indicative of a correlation between its OSGC profile and enhanced infectivity and replication capacity. Sem1 and Sem3 showed similar lytic trends, whereas Sem2 was associated with an early regrowth phase of the bacterial population. MOI is a temporal profiling under specific incubation conditions, in which the effects of changes in the phage-to-bacterial ratio on bacterial growth are examined, with a substantial initial bacterial density (in this case, 0.5 OD_600_) to ensure broad resolution. Many phage-host-environment interactions are involved in this process. For instance, several studies on phages at low MOI levels in the literature report an initial increase in host bacterial growth (Khalatbari-Limaki et al., 2020; Li et al., 2022; Zhong et al., 2024), a trend also observed with our phages (Fig. 1D). If we expand on this example using an MOI of 0.01, which indicates only 1 phage per 100 bacterial cells, the uninfected bacterial cells initially continue to grow, and the phage population is not sufficiently abundant to effectively spread throughout the entire bacterial population. This interpretation is independent of the infective capacity of the phage. In the hours that followed, given the competition between E. coli, which multiply exponentially by dividing in half, and E. coli phages, which replicate by killing bacteria at rates of 80–429 particles per bacterium, the observed decrease in bacterial density was not unexpected. However, it is imperative to acknowledge that factors other than replication capacity and speed of phages must be taken into consideration. Another notable result that was observed in the latter hours, particularly at elevated MOI values in our study, is the rapid increase in bacterial growth following its initial decline. Several explanations may account for this observation: (i) the growth of a resistant subpopulation, (ii) misdirected phage infections due to wall structures released from lysed cells, and (iii) accelerated bacterial growth driven by nutrient release from dying cells. However, the latter two explanations are likely insufficient. In the literature, such outcomes are most commonly attributed to the proliferation of resistant subpopulations (Peng & Yuan, 2018). When examining various MOI-related growth curves generated using different E. coli phages (Khalatbari-Limaki et al., 2020; Zhong et al., 2024), it becomes apparent that the resulting patterns, particularly at extreme MOI values, are often comparable. In line with our findings, Alexyuk et al. (2022) also observed variable bacterial resistance levels and distinct growth patterns depending on the specific phage used (Alexyuk et al., 2022). In summary, MOI profiling reveals highly phage-host-specific interactions, and these dynamics may also vary significantly depending on environmental conditions.

Considering the protein nature of phages, the physiological environment, including changing physiological conditions during infection, or the variability of environmental conditions in biocontrol applications, and even their extremity in some cases, is a noteworthy issue, and phages have been studied in terms of temperature and pH parameters in this regard. All phages studied maintained activity across a broad spectrum, ranging from acidic to basic conditions, specifically within the pH range of 4 to 10. Similarly, when examining studies involving E. coli phages in the literature, a wide variation in pH tolerance is observed, including phages with comparable results (Wintachai et al., 2024), phages capable of remaining active at extremely acidic conditions such as pH 1–2 (Litt & Jaroni, 2017), and phages that exhibit almost minimal activity at pH 10 (Shamsuzzaman et al., 2024). This variability is likely influenced by differences in the composition, frequency, and diversity of proteins found in the capsid and tail structures, and consequently, differences in the net surface charge and other physicochemical properties of the phages. Regarding temperature, it has been observed that E. coli phages retain their activity within the temperature range of 4 °C to 45 °C, while only minimal activity is detected at 65 °C. A review of the aforementioned studies indicates that higher phage activity is typically detected within the lower temperature range (Litt & Jaroni, 2017; Shamsuzzaman et al., 2024; Wintachai et al., 2024). Similar to pH, temperature can influence the external protein topography of the phage, potentially leading to changes in denaturation behavior or adsorption kinetics. In heat-resistant phages, the presence of disulfide bonds between capsid proteins and charged amino acids is known to contribute to the maintenance of structural integrity even at elevated temperatures (Caldeira & Peabody, 2007). Moreover, the capsid surface charge is determined solely by molecular-level factors such as pH, amino acid arrangement, and interactions with the ionic environment (Nap et al., 2014). These observations highlight the necessity of considering such details for an accurate understanding of viral stability and interactions. Above all, the findings from the stability assays indicate that the E. coli phages characterized in this study retain sufficient infective potential under the tested pH and temperature conditions, underscoring their suitability for potential practical applications.

When evaluating the antibacterial effects of phages through MIC analysis, it was generally observed that all phages prevented bacterial proliferation at the two highest titers tested. Another consistently observed finding was that phage activity measured using MHB was superior to that observed with TSB. The choice of culture medium used to assess phage activity can influence both the bacterial growth rate and the replication efficiency of the phage (Ramesh et al., 2019). Additionally, antibacterial activity is influenced by multiple factors, including phage burst size, bacterial division rate, and time-dependent changes in environmental concentrations. These parameters can vary significantly from one phage to another and even between different hosts challenged by the same phage (Kim et al. 2021a, b; Liu et al. 2020). Specifically for phage Sem4, considering that MHB is the standard medium used in antibacterial testing according to antimicrobial susceptibility testing (AST) guidelines (EUCAST reading guide v 5.0, 2024), and that standardized protocols for phage susceptibility testing are currently under development (EUCAST: AST of Phages), its ability to exert minimum bactericidal activity at a concentration as low as 5 × 10¹ PFU/mL demonstrates its potency as a strong antibacterial agent.

The ecological resilience of E. coli across both clinical and environmental settings, including water systems, can be attributed to its remarkable metabolic plasticity, toxin production, and the widespread acquisition of antimicrobial resistance (Dini & De Urraza, 2010; Pormohammad et al., 2019). The role of E. coli, particularly multidrug-resistant strains, in waterborne disease outbreaks linked to contaminated tap or bottled water is increasingly acknowledged. In contrast to conventional chemical disinfectants or filtration systems, both of which are susceptible to clogging, degradation, or operational constraints, phages offer a selective, self-amplifying, and environmentally sustainable alternative capable of targeting both planktonic and biofilm-associated bacterial populations (Elbahnasawy et al., 2021; Kauppinen et al., 2021; Sanchez-Rosario et al., 2024; Zhang et al., 2013). In this study, phage Sem4 demonstrated complete eradication of MDR-Ec in tap water within 24 h, corresponding to an 8-log reduction in bacterial load. Remarkably, this bactericidal effect persisted for 48 h, underscoring the phage’s stability and efficacy under realistic water conditions. These findings position Sem4 as a strong candidate for incorporation into next-generation water treatment strategies, offering a biologically precise and low-maintenance approach to microbial decontamination. It should be noted, however, that since E. coli is often not the sole waterborne contaminant (Kunz et al., 2024) and multiple strains may be present, further studies are needed to evaluate Sem4’s lytic activity against other relevant bacterial species, which would clarify its broader applicability and effectiveness in real-world water treatment settings.

According to the latest classification updates by the International Committee on Taxonomy of Viruses (ICTV), the genus Vectrevirus has undergone a lineage revision in 2024 (Simmonds et al., 2024). In the updated taxonomy, it is now classified within a more refined lineage: Duplodnaviria > Heunggongvirae > Uroviricota > Caudoviricetes > Autographivirales > Autosignataviridae > Molineuxvirinae > Vectrevirus (Taxon Details | ICTV). A total of seventy phages are currently listed as members of the genus Vectrevirus in the NCBI nucleotide database (NCBI Nucleotide, accessed 20 June 2025). In this study, we performed an in-depth bioinformatics analysis of our four phage isolates, which were classified as members of the Vectrevirus genus, and compared them among themselves and with other publicly available Vectrevirus genomes. Although isolated from different geographical regions of Turkey and from different hosts, Sem1, Sem2, Sem3, and Sem4 appear to represent variants of a completely novel species, according to the ICTV species demarcation criteria (Turner, Kropinski et al., 2021). Although the ICTV does not officially recognize the rank of ‘strain’, the term is frequently used informally by researchers to describe closely related genomic variants, particularly in contexts such as discussions of polymicrobial infections caused by multiple phage strains, host specificity, and evolutionary adaptation. In this context, Sem1–Sem4 can be considered distinct strains of the same newly identified species.

Genomic analyses of all phages revealed no antibiotic resistance genes, toxins, or toxin-associated regions. Moreover, no genomic features indicative of a temperate lifestyle were identified. These characteristics collectively suggest that the phages are well-suited for therapeutic applications, as the absence of lysogenic traits and virulence factors is a fundamental prerequisite for their consideration as biological control agents (Dalmasso et al., 2016).

All isolated phages encode a protein highly similar to T7 gp0.3 (Ocr), a B-DNA mimic that protects phage genomes from host type I restriction–modification systems (Isaev et al., 2020). Beyond its canonical role, Ocr has also been shown to inhibit host transcription by binding to RNA polymerase and interfering with sigma factor recruitment (Ye et al., 2020). Asgharzadeh Kangachar et al. (2024) reported Ocr homologs in phages vB_EcoM_SHAK7163 and vB_EcoM_SHAK7704, which they initially described as Vectrevirus members based on their genomic features (Asgharzadeh Kangachar et al., 2024). However, according to the updated NCBI taxonomy, these phages are now placed within the genus Kayfunavirus under the family Autographiviridae. This taxonomic revision suggests that Ocr-like anti-restriction systems may be more broadly distributed across Autographiviridae than previously appreciated. The Ocr homolog in Sem1, Sem2, Sem3 and Sem4 *(*protein id XUY71577.1, XWX23799.1, XWX23472.1, XWX23546.1) shares 100% identity with other Ocr proteins, including those from E. coli phages 6949 (URY99308.1) and EC120 (URC25408.1), further supporting the conserved nature of this anti-restriction function among Vectrevirus phages.

Typically, variations in RBPs, such as tail fibers or tail spikes, are associated with host range adaptation. Phages Sem1, Sem2, Sem3, and Sem4 all share an identical tail fiber proteins, with no observed amino acid substitutions or mutations. This high level of conservation is particularly intriguing given that these phages were isolated using different host strains. Changes in RBPs—whether through amino acid substitutions or the acquisition of novel RBP domains via horizontal gene transfer—are known to influence host specificity significantly (Cho et al., 2025). In their study on phage evolution, Ferriol-González and Domingo-Calap identified vB_Kpl_K44PH129C1 (initially classified as Vectrevirus, now updated in NCBI as Ulipvirus) as a case where RBP mutational hotspots were linked to host-range modulation in a host-diverse environment (Ferriol-González et al., 2024; Ferriol-González & Domingo-Calap, 2024). Similarly, although our isolated phage tail spike proteins show approximately 50% pairwise identity (with 99% query coverage) to the tail fiber protein of vB_Kpl_K44PH129C1 (accession no. CAK6597327.1), no sequence variation was observed among our isolates. This finding suggests that specific positions—such as residue 266 in the tail spike protein—may act as hotspots for adaptive changes in Vectrevirus, and potentially contribute to host range diversification when mutations do occur. A similar phenomenon was reported by Florian Wagenlehner et al., who characterized the phages Escherichia phage vB_EcoP-101101UKE1 (MZ234012), vB_EcoP-UTI89UKE2 (MZ234049), and vB_EcoP-UTI89UKE3 (MZ234050) (Loose et al., 2021). Although these phages are nearly identical at the nucleotide level, as determined by bioinformatics analysis, they exhibit differences in host range. This finding aligns with our results, where phages with highly conserved genomes and tail fiber proteins—consistent with Vectrevirus members, which encode approximately 67 CDS—still exhibit host range divergence. Such cases strongly support the concept of distinct strains within a single phage species, despite the ICTV’s current lack of formal criteria for phage strain designation.

Concerning depolymerases, our isolates share identical Dpo sequences among themselves, which was confirmed by further structural and functional analyses guided by the domain architecture defined by Ferriol-González et al. (2024), who showed high similarity—but not complete identity—to Dpos from other Vectrevirus (Ferriol-González et al., 2024). Despite identical sequences among our isolates, they exhibit subtle amino acid differences (e.g., 99.21%–99.65% identity) when compared to other Vectrevirus Dpos, suggesting evolutionary divergence at the strain level. This pattern of Dpo diversity and its potential link to host specificity is also supported by the findings of Markusková et al. (2024), who described two Vectrevirus phages, vKMB14 and vKMB47 (Markusková et al., 2024), with typical T7-like genome organization (~ 44 kbp) and notable divergence in their tail spike regions. These phages formed clear plaques with halo zones and encoded two distinct Dpos, one of which showed homology to K5 lyase, while the other contained a pectate lyase domain or unique low-similarity sequences, reflecting novel depolymerase specificities. Interestingly, although vKMB14 and vKMB47 shared ~ 89% overall genome identity and with the reference phage VEc3 (NC_047899.1), their host ranges differed (17% vs. 22% of tested strains), further reinforcing the role of tail spike variability—particularly in Dpo domains—in determining host spectrum. Taken together, the depolymerase-related tail spike diversity is a key driver of host adaptation within the Vectrevirus genus, even among phages with otherwise highly conserved genomes. This highlights the importance of Dpo characterization not only for understanding phage biology and taxonomy, but also for informing phage selection in therapeutic applications.

Conclusıon

The increasing prevalence of E. coli infections and foodborne contamination, compounded by the emergence of multidrug-resistant strains, presents significant challenges for treatment and mitigation efforts. Phage-based interventions represent a precise and innovative avenue in addressing the escalating crisis of antibiotic resistance, a global health concern of critical urgency. The phages characterized in this study exhibit inherent bacteriolytic activity and host specificity, underscoring their potential as effective agents in both phage therapy and biocontrol applications. The outcomes of the preliminary assays strongly suggest that these novel phages can be used as antibacterial agents for the treatment and sanitation of E. coli-related infections and contaminations in water. Collectively, these findings highlight the importance of phage–bacteria dynamics and provide a solid basis for advancing phage-focused applications in biotechnology, medicine, and water safety, supporting the development of antibiotic-free antimicrobial solutions.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1

Supplementary Material 2

Supplementary Material 3

Supplementary Material 4