Prophage genomics of carbapenemase-producing Klebsiella pneumoniae from animal-derived food sources Authors: Rana Fahmi Sabala, Akira Fukuda, Masaru Usui, Mohamed Elhadidy Published: 2026 ABSTRACT AbstractCarbapenem-resistant Klebsiella pneumoniae (CRKP) is a globally disseminated pathogen of significant public health concern. Mobile genetic elements, including prophages, frequently drive their multidrug resistance and virulence. However, the contribution of prophages to the genetic landscape of food-derived CRKP remains poorly characterized. The study aimed to characterize the diversity, genetic features, and evolutionary relationships of prophages in CRKP isolates recovered from animal-derived food sources, and to evaluate their potential role in disseminating antimicrobial resistance (AMR) and virulence within the food chain. Whole-genome sequencing was performed on 23 CRKP isolates obtained from foods of animal origin in Egypt. Prophage mining and annotation were conducted to identify and classify prophage elements, detect resistance-and virulence-associated genes, and assess evolutionary relationships through phylogenetic analysis. Eighteen intact prophages were classified as members of the class Caudoviricetes. Based on structural protein analysis, 61.11% exhibited a Siphovirus-morphotype (long, non-contractile tails), while 38.88% were identified as Peduoviridae/Myovirus-morphotype (contractile tails). The blaSHV resistance gene was detected in three incomplete prophages related to PHAGE_Escher_RCS47_NC_042128. Prophage regions also encoded lysozyme-like endolysins, suggesting potential use as lytic agents, and phylogenetic analysis revealed high prophage diversity, likely contributing to CRKP accessory genome evolution, fitness, and virulence. Animal-derived food may serve as a reservoir for CRKP that carry diverse prophages and other mobile genetic elements. Prophage-associated resistance, virulence, and lysis-related genes highlight the potential for dissemination of AMR and pathogenic traits through the food chain, underscoring the importance of surveillance of foodborne CRKP and their mobilome.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-54139-2.Keywords: Carbapenem-resistant Klebsiella pneumoniae, Prophages, Antimicrobial resistance, Food chain, EgyptSubject terms: Computational biology and bioinformatics, Genetics, Microbiology Introduction Klebsiella pneumoniae (K. pneumoniae) is a gram-negative bacterium frequently implicated in nosocomial infections and increasingly recognized for its high-level resistance to different classes of antimicrobials1. A common cause of hospital-acquired pneumonia, K. pneumoniae can also lead to severe, life-threatening infections. Of particular concern, carbapenem-resistant K. pneumoniae (CRKP) has emerged as a critical public health threat due to its resistance to last-resort carbapenem antibiotics. The global spread of CRKP has severely limited treatment options and is often associated with high morbidity and mortality rates1. The continued global emergence and dissemination of CRKP highlights the urgent need for novel surveillance strategies and alternative therapeutic approaches beyond traditional antimicrobials. While K. pneumoniae is a well-recognized nosocomial pathogen, its emerging role as a food contaminant and contributor to food–chain–associated epidemiology is a growing public health concern2. The food supply chain represents a critical route for the transmission of antimicrobial-resistant organisms to humans. Mobile genetic elements (MGEs), including plasmids, transposons, and integrons, play a central role in the rapid dissemination of Antimicrobial resistance (AMR) within bacterial populations by facilitating horizontal gene transfer. In addition to these mechanisms, mobile genetic elements (MGEs) can be mobilized by bacteriophages through transduction or acquired directly from the environment via transformation3. Comprehensive characterization of these elements is therefore essential for elucidating the epidemiological dynamics of carbapenem-resistant K. pneumoniae (CRKP) and informing the development of effective control strategies. Bacteriophages, or phages, are viruses that infect bacteria. Initially proposed as antimicrobial agents in the early 20th century4–6, antibiotics overshadowed their therapeutic use. However, the growing threat of AMR has renewed interest in phage biology, not only for therapeutic purposes but also for understanding mechanisms of horizontal gene transfer, virulence regulation, and bacterial adaptation. Prophages, the integrated form of temperate phages, play a crucial role in bacterial evolution and the dynamics of microbial communities by alternating between lytic and lysogenic cycles, thereby acting as both genetic reservoirs and modulators of host phenotype7. Prophages can carry virulence factors, toxin genes, stress-resistance determinants, and AMR genes, thereby significantly shaping the ecological fitness and pathogenic potential of their bacterial hosts7. However, the prophage content of CRKP strains, particularly those isolated from non-clinical reservoirs such as food of animal origin, remains poorly characterized. This represents a critical knowledge gap, primarily as the food chain may serve as a silent vehicle for the transmission of resistance and virulence traits to human populations. The current study aimed to investigate the genomic landscape of prophages harbored by K. pneumoniae isolates recovered from food of animal origin in Egypt. As foodborne environments are increasingly recognized as important reservoirs for AMR, including isolates carrying mobile genetic elements such as prophages, understanding prophage dynamics in foodborne CRKP strains is essential for assessing the broader public health risks associated with AMR and for informing future interventions targeting phage-mediated gene transfer in both clinical and food-associated settings. Materials and methods Genomic characterization A total of 23 non-clonal CRKP isolates were recovered from 311 food samples of animal origin, including 116 minced meat, 75 diced meat, 29 mutton, and 92 chicken samples collected from 38 local markets in Cairo city, Egypt, between August 2022 and January 2023, as previously reported by Sabala et al.2. Food samples were collected from 38 local markets in Cairo, Egypt, between August 2022 and January 2023. Epidemiological data, such as the sampling date and store location, were recorded, while store names were anonymized. Whole-genome sequencing (WGS) of the isolates was performed as described in the original study2. Libraries were prepared using Nextera XT v2 kits, and sequencing was carried out on the Illumina MiSeq platform with the MiSeq Reagent Kit v3, generating 300-bp paired-end reads2 and were re-analyzed here using a specialized prophage-mining pipeline. Assembled data of the CRKP strains were deposited in DDBJ ( https://www.ddbj.nig.ac.jp/index-e.html ) under the BioProject: PRJDB16935. Prophage in silico detection and characterization The prophages in each CRKP isolate were identified using PHASTER (http://phaster.ca) and phageScope v1.0 (https://phagescope.deepomics.org/), which provide detailed information about the prophage length, region, GC Content, and the number of coding sequences8. PHASTER also classifies detected prophages as intact, questionable, or incomplete based on the length of the phage region and the number of phage-derived genes. Annotation of prophage coding sequences was performed using Pharokka v1.3.2 (https://github.com/gbouras13/pharokka)9, and protein function prediction was conducted using InterProScan (https://www.ebi.ac.uk/interpro/). Prophage taxonomy and lifestyle classification The prophage taxonomy was determined according to the International Committee on Taxonomy of Viruses (ICTV) using phagebox v 2.0 (https://phage.ee.cityu.edu.hk/)10. Furthermore, the prophage defense/antidefense systems were identified using DefenseFinder v3.0.0 (https://github.com/mdmparis/defense-finder)11. The morphotype of the intact prophages was further classified in silico based on the prophage structural head-neck-tail proteins using Virfam v1.1 (https://biodev.cea.fr/virfam/)12. Prophage-associated antimicrobial resistance genes and virulence factors All prophage genomic sequences were screened for antibiotic resistance genes using tools from the Center for Genomic Epidemiology (CGE) (http://www.genomicepidemiology.org/), specifically the ResFinder 4.7.2 database (https://genepi.food.dtu.dk/resfinder). Additional resistance gene identification was conducted using Resistance Gene Identifier (RGI v5.2.1) (https://card.mcmaster.ca/analyze/rgi) and the Comprehensive Antibiotic Resistance Database (CARD v3.2.3) (https://card.mcmaster.ca/analyze) with a minimum threshold of 90% identity and 60% query coverage13. Virulence genes were analyzed using VirulenceFinder 2.0 with a 90% identity and 80% coverage cutoff, selecting Klebsiella as the target species (https://cge.cbs.dtu.dk/services/VirulenceFinder/)14. To validate the integration context of incomplete prophage regions carrying AMR genes, linear genomic comparative maps were generated using the K. pneumoniae CP167055.1 and visualized by Easyfig v2.2.5 (https://mjsull.github.io/Easyfig/) to confirm the presence of phage structural machinery, such as integrase genes, adjacent to the resistance determinant. Phylogenomic analysis of intact prophage sequences Complete prophage sequences were aligned using MAFFT v7.0 with default parameters. The phylogenetic tree of the genomes was subsequently constructed using the maximum-likelihood method in the IQ-TREE tool (Galaxy Version 2.3.6 + galaxy0), and the resulting tree was visualized in iTOL v7 (https://itol.embl.de/itol.cgi)15. A proteome-based clustering strategy was employed to classify the phages into families using the ViPTree server v4.015,16. The analysis was conducted on the phages in the current study and on other closely related dsDNA prokaryotic viruses. Phages with the highest ViPTree tBLASTx scores (SG) were selected to construct a more detailed rectangular proteomic tree. The genome-genome distance was computed by the Virus Classification and Tree Building Online Resource (VICTOR v3.0) (https://ggdc.dsmz.de/victor.php)17 using the Genome-BLAST Distance Phylogeny (GBDP) method18. The analyses were supported by 100 pseudo-bootstrap replicates. The OPTSIL program was used to estimate taxon boundaries at different taxonomic levels19 at recommended clustering thresholds17 and an F value (fraction of links required for cluster fusion) of 0.520. Additionally, a hierarchical clustering algorithm was used to group prophages based on their protein content, and a distance matrix was computed to represent pairwise similarity using Virclust v2.021. Endolysins identification, gene ontology analysis, and functional annotation To assess the presence of lysins within both intact and defective prophages, all prophage sequences were analyzed using Phagescope v1.0 (https://phagescope.deepomics.org/) and InterProScan (https://www.ebi.ac.uk/interpro/). Identified endolysins were annotated with Gene Ontology (GO) using the QuickGO web server (http://www.ebi.ac.uk/QuickGO/). Statistical analysis To assess the association between prophage completeness and the presence of AMR genes, a 3 × 2 Fisher’s Exact Test was performed (comparing Intact, Incomplete, and Questionable types). The distribution of prophage categories was analyzed descriptively, with a proportions test used to compare the prevalence of intact versus incomplete elements. To identify independent predictors of AMR gene carriage, a multivariable logistic regression model was constructed. The model evaluated prophage type as the primary predictor while controlling for potential confounders, including total genome size and Sequence Type (ST). All statistical analyses were conducted in R (v4.3.1), with a p-value < 0.05 considered statistically significant. Results Identification and prevalence of prophages in K. pneumoniae isolates In this study, the genome sequences of the 23 screened CRKP isolates resulted in a total of 50 prophage-like elements. The multilocus sequence typing (MLST) of the 23 CRKP isolates included five different sequence types (STs): ST147, ST383, ST48, ST17, and ST16. Among these elements, 18 were classified as intact, 30 as incomplete, and two as questionable. Intact prophages were identified in 13 (56.52%) of the 23 CRKP isolates, with five isolates harboring two intact prophages. The total number of prophages per isolate ranged from 1 to 4, with most isolates containing two prophages. Genome characteristics of K. pneumoniae prophages The size of the prophages (including intact, incomplete, and questionable) ranged from 6 kb to 50.5 kb, and the number of predicted coding sequences (CDS) ranged from 6 to 76. The GC content ranged from 42.84 to 60.93% with an average of 54.23%. On the other hand, bacterial GC% values ranged from 57.10 to 65.41%, with an average of 56.62% (Supplementary Table 1). For intact prophages, sizes ranged from 43.7 kb to 52.5 kb, and GC content ranged from 51.42% to 54.17% (Table 1). The annotation of these prophages confirmed the typical phage structure modules, including genes for integrase, portal, tail, lysin, terminase, head, and capsid proteins. Table 1The characterization of the intact prophage detected in CRPK isolated from food of animal origin.SampleStrainProphageGC%LengthCDSMorphotypeRelated phagesMinced meatCRKP1CRKP1-653.16%50.2 kb76 Siphovirus PHAGE_Erwini_vB_EhrS_59_NC_048198(21)CRKP1-1051.42%38.9 kb51 Siphovirus PHAGE_Entero_mEp235_NC_019708(16)Diced neatCRKP7CRKP7-152.40%40.2 kb50 Myovirus PHAGE_Klebsi_ST147_VIM1phi7.1_NC_049451(37)CRKP7-553.26%43.7 kb63 Siphovirus PHAGE_Erwini_vB_EhrS_59_NC_048198(20)Minced beefCRKP9CRKP9-151.90%41.3 kb50 Myovirus PHAGE_Klebsi_ST147_VIM1phi7.1_NC_049451(37)CRKP9-853.26%43.7 kb67 Siphovirus PHAGE_Erwini_vB_EhrS_59_NC_048198(20)Chicken meatCRKP10CRKP10-152.10%39.9 kb48 Myovirus PHAGE_Klebsi_ST147_VIM1phi7.1_NC_049451(33)Chicken meatCRKP19CRKP19-454.09%34.6 kb56 Siphovirus PHAGE_Erwini_vB_EhrS_59_NC_048198(15)Diced meatCRKP24CRKP24-453.26%43.7 kb68 Siphovirus PHAGE_Erwini_vB_EhrS_59_NC_048198(19)MuttonCRKP25CRKP25-253.26%43.7 kb66 Siphovirus PHAGE_Erwini_vB_EhrS_59_NC_048198(19)Minced meatCRKP26CRKP26-452.10%39.9 kb49 Myoviru s PHAGE_Klebsi_ST147_VIM1phi7.1_NC_049451(34)Minced meatCRKP27CRKP27-251.90%41.3 kb50 Myovirus PHAGE_Klebsi_ST147_VIM1phi7.1_NC_049451(37)CRKP27-453.16%50.2 kb63 Siphovirus PHAGE_Erwini_vB_EhrS_59_NC_048198(20)Diced meatCRKP29CRKP29-352.10%39.9 kb51 Myoviru s PHAGE_Klebsi_ST147_VIM1phi7.1_NC_049451(35)Minced meatCRKP30CRKP30-554.17%52.5 kb67 Siphovirus PHAGE_Erwini_vB_EhrS_59_NC_048198(19)Minced meatCRKP32CRKP32-153.16%50.2 kb62 Siphovirus PHAGE_Erwini_vB_EhrS_59_NC_048198(20)CRKP32-351.90%41.3 kb51 Myovirus PHAGE_Klebsi_ST147_VIM1phi7.1_NC_049451(39)MuttonCRKP38CRKP38-653.26%43.7 kb68 Siphovirus PHAGE_Erwini_vB_EhrS_59_NC_048198(20)CDS: coding DNA sequence. The characterization of the intact prophage detected in CRPK isolated from food of animal origin. CDS: coding DNA sequence. Classification of K. pneumoniae prophages Among the 50 prophage-like elements detected, 37 were classified according to the ICTV criteria into established families based on genomic architecture, while the remainder were unclassified. The classified prophages belonged to the Caudoviricetes taxonomy. Among the 18 intact prophages, based on the morphotype classification, 61.11% (11/18) of the prophages exhibited Siphovirus morphotype, while 38.88% (7/18) were classified under the Peduoviridae family under the morphological category of Myoviruses. All the Peduoviridae classified prophages were closely related to PHAGE_Klebsi_ST147_VIM1phi7.1, while the unclassified prophages were closely related to PHAGE_Erwini_vB_EhrS_59, except for only CRKP1-10, which was closely associated with PHAGE_Entero_mEp235 (Table 1). Characterization of virulence and AMR genes in K. pneumoniae prophages Screening of 50 prophage genomes revealed three incomplete prophages named CRKP4-6, CRKP27-6, and CRKP35-10, which harbored the blaSHV gene, which encodes a SHV-type β-lactamase enzyme responsible for resistance to β-lactam antibiotics. The frequency of AMR-positive prophage was 6%, compared to 13.04% among the corresponding host genomes. Two of the incomplete prophages were recovered from K. pneumoniae of ST147, while the third one was recovered from K. pneumoniae of ST383. Although AMR genes were exclusively identified within incomplete prophage fragments (n = 3), the association was not statistically significant (p = 0.54). This suggests that, among the analyzed ST isolates, AMR genes are sporadically distributed across the prophage landscape, regardless of the element’s structural integrity. The phage type carrying the AMR gene in the three prophages was related to PHAGE_Escher_RCS47_NC_04212822. The blaSHV gene was located within a 6.2 kb incomplete prophage region, situated 1.2 kb downstream of a phage-specific integras and flanked by structural proteins such as the terminase small subunit, ltnD, and otnK which act as phage-specific machinery. The BLASTn synteny analysis confirmed that this resistance-carrying scaffold is 100% identical across all three isolates and the reference chromosome CP167055.1 (Fig. 1). Notably, this genetic block is located at a known 2.8 Mb insertion hotspot rather than the ancestral ygbI/ygbK chromosomal locus, providing strong evidence that the gene is a mobile prophage-associated cargo rather than a misidentified core feature. The genome comparisons using BLASTn among the three incomplete prophage sequences harboring the blaSHV gene showed > 97% identity while they recovered from CRKP isolates of different STs (Fig. 1). Gene ontology analysis linked the prophage region to GO:0006313 (DNA integration) and GO:0008800 (β-lactamase activity), reflecting the co-localization of the resistance gene with the integration machinery. Comparative genomic analysis of the blaSHV carrying prophages in carbapenem-resistant K. pneumoniae (CRKP) isolates: Linear synteny comparison of the mobile genetic scaffold containing blaSHV across three foodborne CRKP isolates (CRKP4-6, CRKP27-6, and CRKP35-10) and reference CP167055.1. Arrows represent open reading frames (ORFs) and indicate transcriptional direction. The conserved genetic neighborhood includes recF (DNA repair), the resistance gene blaSHV (a key antibiotic resistance gene), the regulatory gene glpR (glycerol-3-phosphate repressor), and phage-associated proteins ltnD (phage tail-related), otnK, and otnC. Gray shaded blocks indicate 99%–100% nucleotide identity (BLASTn). Although no classical virulence genes were detected, several fitness and regulatory elements were identified. (Supplementary Table 2). The most prevalent was the transcriptional regulator TraR/DksA, which mediates stress-response gene expression23,24. These elements often co-occurred with LexA and CII family regulators, suggesting complex intra-prophage regulation. Additional features included crossover junction endodeoxyribonuclease RusA, tyrosine-type recombinase/integrase, type II toxin-antitoxin systems (RelE/ParE), and fimbria/pilus-associated periplasmic chaperones (Supplementary Table 2). Prophage CRKP28-9 harbored Type Six Secretion system D (tssD), a core component of the Type VI Secretion System (T6SS), which is a needle-like system that injects the bacteria with harmful proteins (effectors), such as pore-forming toxins and nucleases, and promotes bacterial survival in other ways, including facilitating acquisition of resources. CRKP24-8 encoded sapC and sapD, conferring resistance to host antimicrobial peptides. Acid-protective chaperones evgA and hdeB were found in CRKP30-5. Anti-CRISPR proteins were present in 20% (10/50) of prophages, 50% (9/18) of intact versus 3.3% (1/30) of incomplete prophages. All anti-CRISPR-positive intact prophages belonged to the Siphovirus-morphotype prophages, while the incomplete CRKP10-3 encoded Tad2, an anti-defense protein that inhibits the Thoeris antiviral system. Characterization of prophage-encoded lysis proteins in K. pneumoniae Most intact prophages encoded phage lysis-associated proteins, including Kil protein for bacterial septation inhibition, lysozyme, holin, and Rz-like spanin (Table 2). Kil protein delays host-cell division, thereby optimizing phage replication. The combined action of holin, lysozyme, and spanin facilitates host cell lysis and progeny release. Bioinformatic characterization of the lysis modules in intact prophages revealed endolysins belonging to the Glycoside Hydrolase family 24 (GH24). These Lambda-type endolysins possess muramidase catalytic activity (GO:0003796), specifically targeting the β-(1→4) glycosidic bonds in bacterial peptidoglycan. The regulatory control of these lysis modules is further supported by the identification of antitermination Q-gene homologs within several intact prophages. The Q-protein acts as a master transcriptional switch; it modifies RNA polymerase to bypass termination sites, thereby enabling expression of the late operon, which contains the aforementioned lysis and structural genes. Table 2The lytic features of the intact prophage recovered from the 23 CRKP isolates. Lysis genes are stated Lysozyme (L), Lytic transglycosylase (GH), Holin (H), Spanin (S), and Kil protein (K).Phage nameLysis genesOther lysis factorsEndolysin typeCRKP1-6L, H, SEndolysin T4 typeCRKP1-10GH, H, S, KNDCRKP7-1L, H, SLysB family, TCT, GPOEndolysin lambda typeCRKP7-5L, H, S, KEndolysin T4 typeCRKP9-1L, HLysB family, TCT, GPOEndolysin lambda typeCRKP9-8L, H, S, KEndolysin T4 typeCRKP10-1L, HLysB family, TCT, GPOEndolysin lambda typeCRKP19-4L, H, S, KFTHFDEndolysin T4 typeCRKP24-4L, H, S, KEndolysin T4 typeCRKP25-2L, H, S, KEndolysin T4 typeCRKP26-4L, HLysB family, TCT, GPOEndolysin lambda typeCRKP27-2L, HLysB family, TCT, GPOEndolysin lambda typeCRKP27-4L, H, S, KEndolysin T4 typeCRKP29-3L, HLysB family, TCT, GPOEndolysin lambda typeCRKP30-5L, H, S, KFTHFD, hydolyze UAGPP,Endolysin T4 typeCRKP32-1L, H, S, KEndolysin T4 typeCRKP32-3L, HLysB family, TCT, GPOEndolysin lambda typeCRKP38-6L, HLysB family, GPOEndolysin lambda type The lytic features of the intact prophage recovered from the 23 CRKP isolates. Lysis genes are stated Lysozyme (L), Lytic transglycosylase (GH), Holin (H), Spanin (S), and Kil protein (K). Comparative genomics of K. pneumoniae prophages Among the 18 intact prophages, close evolutionary relationships were observed, with clusters sharing functionally similar gene sets (Fig. 2A). Groups such as CRKP7-5, CRKP9-8, and CRKP25-2, also among CRKP1-6, CRKP27-4, and CRKP32-1, as well as among CRKP9-1, CRKP27-2, and CRKP32-3, showed high genomic similarity (Fig. 2B–D). A comparative genomic analysis of 18 intact prophages found in 13 CRKP isolates. Panel A: A phylogenetic tree illustrating the evolutionary relationships among the prophages, alongside a heatmap showing the gene content across various functional categories. Panels B–D: These panels present detailed comparative synteny maps of gene order and regions of sequence homology (indicated by colored ribbons) between different prophage clusters. The color of the arrows in these maps corresponds to the function of each gene, while the ribbon colors represent the percentage of sequence identity. OPTSIL-based clustering revealed seven clusters, representing three genera within the Myovirus- and Siphovirus-morphotype prophages. These findings were consistent with VirClust results, which showed that each cluster exhibited high genetic similarities and functional gene diversity (Fig. 3A). Heatmaps confirmed distinct clusters of closely related prophages with shared ancestry (Fig. 3B). Depicts the phylogenetic relationships and genomic distances among 18 prophages identified in . K. pneumoniae. Panel A presents a clustering tree, illustrating two distinct clades corresponding to the Peduoviridae family of Myovirus-morphotype and unclassified Siphovirius-morphotype prophages. The color-coded bars below the tree indicate the number of putative prophage-encoded proteins associated with each prophage. Panel B displays a heatmap visualizing the intergenomic distances between the prophages. The color gradient represents the degree of genomic similarity, with darker shades indicating higher similarity. Proteome-based phylogeny, using VipTree, compared the intact prophages genomes against ∼4,800 phages, showing 11 prophages were closely related to the unclassified Siphovirus-morphotype prophages, and 7 to the unclassified phage of the Myovirus-morphotype prophages within the host group Pseudomonadota (Fig. 4). Viral proteomic trees of the intact phages: (A) Circular proteomic tree based on genome-wide similarities among the current study phages (red pentacle) and closely related reference phage genomes. (B) rectangular proteomic tree of the current study phage and the reference phages with the highest VipTree SG scores. Phylogenetic comparison with Egyptian K. pneumoniae prophages available on the PATRIC database revealed that CRKP1-10 closely resembles vB_Kpn_ZCKp20p (OP373729), a therapeutic phage with potent lytic and anti-biofilm activity (Fig. 5). Genomic analysis showed these prophages encoded full suites of genes required for phage replication, regulation, and host lysis using phagescope (Supplementary Fig. 1). The phylogenetic tree for the intact prophage recovered from the current study and the Klebsiella phage retrieved from previous studies in Egypt (available on the PATRIC website). Discussion This study provides a comprehensive in silico analysis of prophages identified in 23 CRKP isolates recovered from food of animal origin. Our genomic characterization revealed 18 intact, 30 incomplete, and 2 questionable prophages, with every isolate harboring at least one prophage. These findings underscore the ubiquity of prophage elements in K. pneumoniae, consistent with prior studies reporting that all clinical isolates contain prophages25. The dissemination of prophages in these food-origin isolates mirrors clinical trends, yet their ecological roles appear distinct. In the food chain, prophages likely function as stable genomic reservoirs that prioritize host survival under environmental stressors, such as temperature fluctuations or preservatives, over the rapid lytic cycles common in clinical settings. The identification of incomplete, defective fragments suggests a strategy of long-term persistence rather than active mobilization. By maintaining functional cargo within these stable elements, K. pneumoniae avoids the metabolic fitness costs of viral replication, allowing these genetic traits to remain fixed within the lineage as it moves from animal reservoirs through the food supply. The presence of multiple prophages in some isolates further supports the polylysogenic nature of K. pneumoniae, reflecting the complex evolutionary interplay where prophages significantly shape the host’s genomic architecture and evolutionary dynamics26. The GC content of the identified prophages (range 42.84–60.93%) showed a downward shift compared to that of the host bacterial genomes (range 57.10–65.41%). This compositional heterogeneity is a hallmark of recent acquisition from external bacterial sources. Because prophages typically undergo amelioration to match the host’s mutational bias over time, the presence of low-GC ‘genomic islands’ suggests that these elements are relatively recent additions to the host pangenome rather than ancient, highly adapted residents. The predominance of incomplete prophages was expected, as intact prophages are often subject to selective pressure or genetic degradation by their bacterial hosts, leading to their eventual loss from the genomes27–29. Polylysogeny in these foodborne isolates represents a significant public health concern, as the accumulation of multiple prophages increases the potential for the synergistic dissemination of diverse AMR and fitness genes. By acting as a combinatorial reservoir for these traits, polylysogenic strains facilitate the emergence of robust, multidrug-resistant (MDR) populations within the food supply, which can ultimately be transmitted to human populations, complicating clinical treatment strategies. This high prophage burden specifically drives genomic plasticity by providing a flexible accessory gene pool that enables rapid acquisition of niche-specific survival tools essential for persisting within the food chain reservoir. Of particular concern, genomic analysis revealed the presence of blaSHV gene encoding beta-lactam resistance in three incomplete prophages. To date, prophage AMR genes have not been reported in K. pneumoniae isolates from Egypt25,30. However, prophages harboring multiple AMR genes conferring resistance to beta-lactams and carbapenems were identified in hypervirulent K. pneumoniae31. Furthermore, AMR genes (ARGs) conferring resistance to macrolides and aminoglycosides were carried on prophages32–34. The linear genomic mapping confirmed that the blaSHV gene was flanked by the phage-associated gene ltnD rather than the typical chromosomal neighbor ygbI, providing strong evidence that these genes have been captured by prophage sequences. In the current study, although the prophages carrying blaSHV are incomplete and likely non-inducible, they serve as stable genomic reservoirs for blaSHV gene and remain potential mobile elements encoding AMR gene35. However, their role as ‘potential mobile elements’ should not be overlooked; while defective in the viral sense, such prophage remnants can be mobilized via alternative horizontal gene transfer mechanisms. These fragments may be sequestered within larger Integrative and Conjugative Elements (ICEs) or mobilized during recombination events with other MGEs, thereby facilitating the dissemination of blaSHV across the K. pneumoniae landscape despite the lack of a functional lytic cycle. These findings suggest that even defective viral elements can act as stable, long-term reservoirs for resistance, facilitating the dissemination of blaSHV through alternative horizontal transfer mechanisms and thereby increasing the risk of multidimensional AMR transmission from the food chain to clinical environments. Notably, the identification of a near-identical blaSHV-carrying prophage region across divergent STs (ST147, ST16) provides robust genomic evidence for the horizontal dissemination of antimicrobial resistance within the food chain reservoir, rather than mere clonal expansion. Although classical virulence genes, such as toxins and adhesins, were absent, several functional genes associated with bacterial fitness and regulatory control were identified within the prophages. These included transcriptional regulators like TraR/DksA, LexA, and CII, which are known to modulate stress responses and bacterial fitness 23,24. Furthermore, enzymes such as the crossover junction endodeoxyribonuclease RusA and tyrosine-type recombinases, which contribute to genome stability, were also detected. The integration of T6SS components within the CRKP28-9 prophage suggests a lysogenic conversion event that enhances host fitness. By providing a contact-dependent killing mechanism, the prophage may confer a competitive advantage to the K. pneumoniae host in polymicrobial environments, facilitating niche establishment and improved resource acquisition by inhibiting competing bacteria. Collectively, these findings indicate that prophage regions may harbor genes associated with bacterial fitness and regulatory functions. While such elements could potentially contribute to bacterial adaptation, the present study was limited to genomic prediction and did not assess gene expression or phenotypic effects. Therefore, the functional contribution of these prophage-associated genes to K. pneumoniae pathogenicity remains to be determined. This highlights a potential indirect role of prophages in supporting the food-borne transmission of K. pneumoniae by promoting environmental resilience and colonization capacity. These findings underscore the importance of monitoring prophage content in foodborne strains, particularly in the context of One Health and the dissemination of AMR. This is especially relevant as MDR bacteria are increasingly isolated across the One Health continuum, spanning animals, the environment, food, and plants, further highlighting the interconnected nature of resistance dissemination36. Our results align with recent reports emphasizing the role of environmental and animal reservoirs in the persistence and global spread of clinically significant pathogens37. Consequently, characterizing the prophage landscape is essential for understanding how K. pneumoniae persists within the food chain and ultimately transitions into human clinical environments. To minimize the limitations arising from the functional degradation of defective prophages, subsequent analysis focused on intact prophages. Classification of the intact prophages revealed a predominance of unclassified Siphoviridae-morphotype (61.11%) and Peduoviridae family under the morphological category of Myoviruses (38.88%), both belonging to the order Caudovirales. These findings are consistent with previous reports showing the predominance of Siphoviridae and Myoviridae among K. pneumoniae isolates. For instance, a study of clinical isolates identified 62 Myoviridae prophages (59.6%) and 40 Siphoviridae prophages (38.5%)25, reinforcing our observations. The predominance of these families among intact prophages from food sources reflects a conserved prophage composition across clinical and non-clinical settings. This suggests that food may serve as a reservoir for phage elements that contribute to the genetic diversity and adaptability of K. pneumoniae, with potential implications for cross-environmental transmission. The identification of putative GH24 and T4-type endolysins, along with antitermination Q-gene homologs, provides a comprehensive view of the lytic potential of these foodborne prophages. While the efficacy of endolysins against Gram-positive pathogens such as methicillin-resistant Staphylococcus aureus (MRSA) is well documented38–40, we recognize the structural barrier posed by the outer membrane in Gram-negative K. pneumoniae. Rather than proposing these specific enzymes as ready-to-use clinical therapeutics, our findings identify the food microbiome as a natural genetic library of diverse lytic machinery. The presence of these homologs suggests that these prophages are not only structurally intact but also maintain the regulatory machinery necessary for active lytic induction. This implies that under conditions of bacterial stress, such as exposure to certain antibiotics, these prophages could be triggered to lyse the host cell and potentially facilitate the horizontal transfer of fitness factors or AMR genes to the surrounding K. pneumoniae population. Experimental induction assays would be required to confirm prophage activation and the functional activity of these enzymes. Phylogenetic analysis of the evolutionary trajectories of K. pneumoniae prophages. These findings highlight the potential contribution of prophages to bacterial virulence, fitness, and pathogenicity. This was supported by the evidence that the MLST analysis of the 23 isolates identified diverse sequence types (ST17, ST48, ST147, ST383)2. The presence of near-identical prophages across these diverse STs suggests horizontal dissemination rather than vertical clonal inheritance. The detailed characterization of prophages in these isolates enhances our understanding of their genetic variability and underscores their potential role in the dissemination of AMR and virulence traits, reinforcing the need for integrated surveillance strategies that account for prophage dynamics in both clinical and foodborne contexts. While this study provides important insights into the prophage landscape of foodborne K. pneumoniae, several limitations must be considered. First, although the 23 isolates represent five distinct Sequence Types (STs), they may reflect specific local clonal lineages rather than the total genetic diversity of K. pneumoniae within the global food supply. Second, reliance on in silico tools like PHASTER is a primary constraint; while the identification of “Intact” prophages complements GH24 endolysins and the Q-gene antiterminator, suggesting viability, computational prediction does not always equate to functional induction or a successful lytic cycle in a laboratory setting. Consequently, these elements might exist as “pseudo-prophages” with impaired mobilization potential. Finally, the relatively small sample size may not fully account for the impact of clonal expansion or capture rare prophage-mediated traits, necessitating future large-scale genomic and functional studies to validate these findings across a broader phylogenetic spectrum. Conclusion This study reveals a diverse repertoire of prophage-encoded genes involved in lysis, replication, and regulation in CRKP isolates recovered from animal-derived food sources, highlighting the potential contribution of prophages to the genomic plasticity of this pathogen. The identification of resistance- and fitness-associated genes within prophage regions suggests that prophages may represent an additional genomic reservoir of adaptive traits in foodborne CRKP strains. However, several limitations should be acknowledged. This study was based solely on in silico genomic analyses, and no experimental validation was conducted to assess prophage inducibility, gene expression, or prophage-mediated horizontal gene transfer. Therefore, the functional roles of the identified prophage-associated genes, including endolysins and resistance determinants, remain to be experimentally validated. Future research should focus on experimental characterization of prophage functionality, including induction assays under stress conditions, such as mitomycin C exposure or UV irradiation, to assess prophage activation and lytic potential. Additionally, investigations into prophage-mediated transduction and gene expression will be necessary to determine their roles in the dissemination of antimicrobial resistance and bacterial adaptation. Broader genomic surveillance and metagenomic studies along the food production chain may further clarify the ecological and epidemiological significance of prophages in the spread of antimicrobial resistance within a One Health framework. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Abbreviations Antimicrobial resistance Antimicrobial resistance genes Comprehensive antibiotic resistance database Coding sequences Center for genomic epidemiology Carbapenem-resistant Klebsiella pneumoniae DNA Data Bank of Japan Deoxyribonucleic acid Double-stranded Deoxyribonucleic acid Genome-BLAST distance phylogeny Guanine-cytosine Gene ontology Integrative and conjugative elements International Committee on Taxonomy of Viruses Interactive tree of life Multidrug-resistant Mobile genetic elements Multilocus sequence type Methicillin-resistant Staphylococcus aureus N-acetylglucosamine N-acetylmuramic acid National library of medicine Picoampere Phage search tool Type VI secretion system Virulence factors Virus classification and tree building online resource Whole-genome sequencing World Health Organization Extensively drug-resistant Author contributions RFS. AF, MU, and ME designed the overall project, analyzed the results, and prepared the manuscript. RFS drafted the manuscript. RFS. AF, MU, and ME collected and analyzed the results. RFS. AF, MU, and ME reviewed and revised the final version of the manuscript. Funding This work was funded by a full postdoctoral fellowship from the Ministry of Higher Education of the Arab Republic of Egypt. Data availability Assembled data of the CRKP strains were deposited in DDBJ ( [https://www.ddbj.nig.ac.jp/index-e.html](https:/www.ddbj.nig.ac.jp/index-e.html) ) under the BioProject: PRJDB16935. Declarations Competing interests The authors declare no competing interests. Consent for publication I Mohamed Elhadidy the corresponding author of the manuscript, hereby provide consent for the publication of our manuscript entitled “Prophage Genomics of Carbapenemase-Producing Klebsiella pneumoniae From Animal-derived Food Sources” in Scientific Reports. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Masaru Usui, Email: usuima@rakuno.ac.jp. Mohamed Elhadidy, Email: melhadidy@zewailcity.edu.eg. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Data Availability Statement Assembled data of the CRKP strains were deposited in DDBJ ( [https://www.ddbj.nig.ac.jp/index-e.html](https:/www.ddbj.nig.ac.jp/index-e.html) ) under the BioProject: PRJDB16935.