Characterization of two novel lytic bacteriophages having lysis potential against MDR avian pathogenic Escherichia coli strains of zoonotic potential

Avian pathogenic E. coli (APEC) is associated with local and systemic infections in poultry, ducks, turkeys, and many other avian species, leading to heavy economical losses. These APEC strains are presumed to possess zoonotic potential due to common virulence markers that can cause urinary tract infections in humans. The prophylactic use of antibiotics in the poultry sector has led to the rapid emergence of Multiple Drug Resistant (MDR) APEC strains that act as reservoirs and put human populations at risk. This calls for consideration of alternative strategies to decrease the bacterial load. Here, we report isolation, preliminary characterization, and genome analysis of two novel lytic phage species (Escherichia phage SKA49 and Escherichia phage SKA64) against MDR strain of APEC, QZJM25. Both phages were able to keep QZJM25 growth significantly less than the untreated bacterial control for approximately 18 h. The host range was tested against Escherichia coli strains of poultry and human UTI infections. SKA49 had a broader host range in contrast to SKA64. Both phages were stable at 37 °C only. Their genome analysis indicated their safety as no recombination, integration and host virulence genes were identified. Both these phages can be good candidates for control of APEC strains based on their lysis potential.

In developing countries like Pakistan, population density has placed a great pressure on animal rearing practices. To have maximum production in small places with poor housing, antibiotics are frequently prescribed, often in sub therapeutic doses to avoid infection and as growth promoters 11 . That results in the origin of multiple drug resistance (MDR) bacterial strains in animal sector. Presence of such MDR strains in food chain that possess common virulence markers as human Uro-pathogenic E. coli strains (UPEC) can lead to greater exposure of human population to these MDR strains with probable zoonotic potential and pose threat to human health resulting in regular outbreaks leaving little choice for physicians to treat them 12 To control such MDR variants to non-infectious levels phage therapy has been a subject of interest from a long time 13,14 . Bacteriophages are natural killers of bacteria and are host specific unlike broad spectrum antibiotics that affect commensal bacterial populations along with pathogenic ones. They are abundant in almost every habitat 15 . In last two decades the use of lytic bacteriophages to control infections in burn sepsis patients and wound infections have given a new hope to control MDR bacteria [16][17][18] . Several studies have reported isolation and characterization of phages against APEC isolates as well 19,20 A recent study has also reported potential of phages to reduce APEC load in vivo 21 However, all this success majorly depends on discovery and detailed characterization of new phage species that can specifically target pathogenic bacteria. Phage must possess few basic characteristics for use as biocontrol or therapeutic agent such as being exclusively lytic (carrying no genes for lysogeny and recombination), possess good efficacy and safety, and should have pH and temperature stability for normal storage. Preferably they should be completely sequenced, and their genomes should be evaluated for presence of toxins and resistance markers. In addition, phages must have a good virulence potential against target bacteria. Testing a complete list of desired phage properties can be limiting due to financial and time constraints however, certain basic set of evaluations must be performed before using any phage [22][23][24] . In this context we designed this study to isolate, characterize and evaluate virulence potential and check stability of two novel lytic bacteriophages against MDR strain of E. coli of poultry origin. We also characterized phage genomes for presence of any toxins or virulence genes.
One step growth curve and stability of phages. Escherichia phage SKA49 and SKA64 were isolated from poultry tissue samples (broiler chicken) previously tested positive for presence of Escherichia Spp., using Escherichia coli strain QZJM25 (Gene Bank Acc. No OK086691) as host. Phages produced clear zones of lysis in titration indicating their lytic nature (Fig. 1A). Both SKA49 and SKA64 grew best at 37 °C giving highest titer of 1.8 × 10 9 pfu / ml and 1.42 × 10 8 respectively, however they were not infective at higher temperatures as no plaques were observed after incubation for 1 h at 60 °C, and 80 °C (Fig. 1B). SKA49 had better pH tolerance than SKA64, as it survived well at pH5 (5.83 × 10 5 pfu/ml) and pH9 (3.93 × 10 6 pfu/ml) in addition to pH7 (1.79 × 10 9 ) albeit with significant (p = 0.0001) drop in titer. SKA64 on other hand was infective only at pH7 (2.8 × 10 9 ) and no plaques were observed after incubation at pH5 and pH9. Neither phage survived at pH 12 (Fig. 1C). The burst size and latent period of both phages was calculated using a one-step growth curve as illustrated in materials and methods. The latent time for SKA49 was 35 min whereas it produced 56 virions /cell (± 7) (Fig. 1D). Whereas SKA64 has a 30 min latent period and burst size of 33 virions/cell (± 5) (Fig. 1E)   Genome analysis of Escherichia phage SKA49. SKA49 possesses a double stranded DNA genome of 51.548 Kilo base pairs with a GC content of 44%. When compared in nucleotide database BLASTn, SKA49 shows nucleotide homology to other members of genus Carltongylesvirus viruses. However, the homology was less than 95% to other Carltongylesvirus phages therefore it was classified as new species as per ICTV rules and named as Escherichia phage SKA49. When analyzed in VIRIDIC software maximum nucleotide homology of SKA49 was observed for Escherichia phage vB_EcoM_Bp10 (Bp10, 85.4%) then for Escherichia phage Mangalitsa (Mangalitsa, 84.6%), Escherichia phage flopper (flopper, 84.5%), Escherichia phage vB_EcoP_Bp7 (Bp7, 84.1%), Escherichia phage ST32 (ST32, 84%), and Escherichia phage phiEcoM-GJ1 (GJ1, 81.5%). All these phages are either members of genus Carltongylesvirus or unclassified Carltongylesvirus of sub-family Cleopatravirinae, in family Chaseviridae (Fig. 3). A cluster table generated by VIRIDIC software indicating SKA49 species and genus placement along with its close homologs is given in supplementary Table S6A. SKA49 exhibits a modular organization as proteins involved in similar functions were clustered together. A genome map of SKA49 is given in Fig. 4. For ease of description genome is divided into five modules [1][2][3][4][5] . In total SKA49 genome encodes 72 Open Reading Frames (ORFs). Each ORF is labelled according to its description in supplementary Table S7, and it is color coded to indicate putative function. Out of 73 ORFs, 64 ORFs started with ATG start codon, 6 proteins including one hypothetical protein (ORF 8,29,35,49,63 and 64), started with GTG start codon . Additionally, two proteins including dNMP kinase (ORF 42) and a hypothetical protein (ORF 54), used TTG as start codon. They all specified methionine as start codon. ICTV core gene analysis as well as CoreGenes 5.0 analysis identified 52 ORFs out of 73 as core genes in SKA49 specific for the entire genus of Carltongylesvirus when compared to its close homologs in NCBI (supplementary Table S8).
Module 1 included majority ORFs encoding hypothetical proteins with unknown functions (ORFs 2-30, Fig. 4, Table S7). These proteins were genus specific as they exclusively matched genus Carltongylesvirus or unclassified Carltongylesvirus with no other hits in BLASTn. These proteins are small, ranging from 42 AA (ORF   Fig. 4, Table S7). Nucleotide sequence of entire SKA49 tail module has highest similarity to phage Mangalitsa (94% coverage and 90% similarity). This module included three distinct sections. Section one included ORFs involved in tail tube formation. The second section included genes involved in base plate formation and tail tube initiation whereas third section included genes involved in tail fiber formation. The first two sections (tail tube and base plate Table 1. SKA49 and SKA64 host range as tested on field isolates of Escherichia coli. E. coli p (isolates of poultry origin), E. coli H (isolates of human origin), *Host strain for SKA49 and SKA64 isolation, (+ + + +) clear lysis, (+ + +) Too numerous to count plaques, TNTC, (+ +) countable plaques ( +) light zone, (-) no lysis. www.nature.com/scientificreports/ ORFs) were conserved with genus specific homology however the third section tail fiber related proteins were less conserved. ORF 70, a putative tail fiber protein appears to be a chimera between the two phages. Nucleotide sequence of 5' end of this tail fiber gene (ORF 70) had ~ 70% homology (albeit with low query coverage 65%) with Escherichia phage flopper whereas the 236 nucleotides at the 3′ end had high homology (84%) with Escherichia phage Mangalitsa. Roughly 400 nucleotide sequence in between these two was unique to SKA49 (supplementary Figure S2). The 5′ end of ORF 70 protein has an R type Pyocin knob domain a putative high molecular weight bacteriocin involved in binding of core LPS of host bacterial strain. This domain was present only in members of genus Carltongylesvirus, however maximum homology of SKA49 domain was observed with Escherichia phage flopper (78% amino acid homology). The comparison of tail fiber is given in supplementary Figure S2. SKA49 has two putative cell wall hydrolases distributed in two modules of the genome (Fig. 3). A putative L-alanyl-Dglutamate peptidase hydrolase (ORF 22) and a second cell wall hydrolase was present in holin-hydrolase combination (ORF 70 and 71) and a putative spannin protein (ORF 31) (supplementary Table S7). A comparison of SKA49 proteins with its close homologs is given in Fig. 5.

Genome analysis of Escherichia phage SKA64. Escherichia phage SKA64 is a member of the subfamily
Stephanstirmvirinae genus Phapecoctavirus. It has 152.401 kb genome with 39% GC content. When compared in BLASTn SKA64 was closely related to members of genus Phapecoctavirus however this homology was less than   Table S7, are given at the top of each ORF.   Table S6B). It showed maximum homology with Escherichia phage PNJ1809-36 (94.3%), followed by Escherichia phage BI-EHEC (93.6%). A heat map for percentage nucleotide homology between close homologs of SKA64 generated by using VIRIDIC software is given Fig. 6. The genome map of SKA64 is given in Fig. 7.The genome has 275 ORFs that are color coded according to their putative function. Only those ORFs are labelled by a number whose function is known. Out of total 275 ORFs vast majority are hypothetical proteins (188 ORFs) having BLASTn homology with phage proteins of unknown functions (color orange). Out of remaining 87 ORFs having BLASTn homology with characterized proteins, 16 ORFs were involved in tail formation (color blue) and 8 ORFs in capsid formation (color black) whereas others were conserved proteins involved in DNA metabolism and replication (color red). SKA64 has a large pool of tRNA genes including two for Methionine, 3 for Selenocysteine and one each for Isoleucine, Proline, Glutamine, Glycine, Threonine, Asparagine, Tyrosine, Serine, and Arginine. These tRNA genes are clustered together after large subunit of phage terminase. The presence of these tRNAs indicate codon usage autonomy of SKA64 (color olive green).
All ORFs involved in tail formation (colored blue, Fig. 7) are conserved throughout the genus except for SKA64 tail spike protein (ORF 253, 2550 bp, 849AA). This protein has no nucleotide or protein homology with any member of genus Phapecoctavirus or even sub-family Stephanstirmvirinae. The closest homolog of this tail fiber protein is Escherichia phage Pisces and Penshu1, both from genus Kayfunavirus from another family. SKA64 has a non-contractile tail sheath like other members of this genus. The nucleotide sequence of capsid genes is conserved in the genus. SKA64 has a large pool of genes involved in replication, transcription, . The genome map has labels for all these genes that correspond to their description in supplementary  Table S8. SKA64 utilizes a spore cortex-lytic enzyme of hydrolase 2 superfamily and is predicted by HHpred to be a lytic transglycosylase (https:// toolk it. tuebi ngen. mpg. de/ tools / hhpred). SKA64 has a recombinase family protein (ORF 101, supp. Table S8) that is in nature a resolvase responsible for keeping phage DNA concatemers as single genome. SKA64 had homology with only phage genomes and no prophage matches were found which indicates it does not integrate in host genome.
Suitability of SKA49 and SKA64 as probable therapeutic/biocontrol agents. One of the important characteristics required for any phage for putative therapeutic/ biocontrol use is that it must be exclusively lytic, possessing no integration / recombination genes in its genome. Moreover, it must not have any virulence genes or antimicrobial resistance markers. To test the suitability of SKA49 and SKA64 as probable candidates for therapy/ control of bacteria they were subjected to Phage Leads software analysis. The absence of integration and recombination genes was indicated by Phage Leads software predicted that both phages may be following lytic pathway. Both SKA49 and SKA64 had no virulence and antimicrobial resistance genes as indicated by Phage Leads software analysis.
Phylogenetic analysis of SKA49 and SKA64. Genome-BLAST Distance phylogenomic (GDBP) tree was produced using close homologs of SKA49 and SKA64. The comparative analysis and accession numbers of these taxa for both phages are given in supplementary Tables S9A and S9B respectively. The phylogenomic GBDP tree inferred using the formulas D0 for SKA49 is given in Fig. 9A. It yielded average support of 7%. The branch lengths of the resulting VICTOR trees are scaled in terms of the respective distance formula used. For SKA49 The OPTSIL clustering yielded twelve species clusters. All except one (OK148439) E. coli phages clustered together, and all Yersinia phages clustered together. OK148439 appeared as a basal group to Yersinia phages. The SKA49 phage (OL741059) appeared next to the basal group (EF460875) of the E. coli phage cluster. This analysis identified 12 species clusters and SKA49 was closest to Escherichia phage Mangalitsa and GJ1. To generate proteomic tree for SKA49 ViPTree software was used to compare genome wide sequence similarities using tBLASTx. All phages having genome similarity score between 1-0.02 for SKA49 (30 phages) were included in phylogenetic analysis (supplementary Table S10A). It further validated placement of SKA49 in genus Carltongylesvirus closest to GJ1 and Mangalitsa. This tree placed Escherichia phage Mangalitsa, Escherichia phage ST32 and Escherichia phage flopper in one cluster recognizing SKA49 as a separate species in the same genus. Tree represents SKA49 relationship with close homologs and outliers (Fig. 9B).  Table S9, are given at the top of those ORFs whose putative function is known. www.nature.com/scientificreports/ For SKA64 the GDBP tree was inferred using D0 formula yielding an average support of 2%. OPTSIL clustering produced fourteen species clusters for SKA64 identifying it as a separate species in the genus (Fig. 9C). When nucleotide similarities were compared SKA64 has closest homology with Escherichia phage phAPEC8. SKA64 proteome tree was generated using 39 phages showing genome similarity range of 1-0.07 in a genome similarity table produced by ViPTree software (supplementary Table S10B). According to proteomic tree SKA64 was more closely related to Klebsiella phage ZCKP1 and Escherichia Phage ESCO13. Both analysis place SKA64 in genus Phapecoctavirus as a new species (Fig. 9D).
Furthermore, we also generated a maximum likelihood tree using amino acid sequence of one ortholog identified in CoreGenes analysis for both phages. This tree was produced using TreeDyn software. For SKA49 Major Capsid gene was used. Software produced and alignment of 336 amino acids. The best fit model on the data was found to be WAG + I. Tree produced is given in Fig. 10A. It placed SKA49 closest to Escherichia phage Mangalitsa, further strengthening our analysis.
For SKA64 Tail fiber protein was selected as the ortholog and a maximum likelihood tree was produced using an alignment of 959 amino acids. The tree places SKA64 closest to Klebsiella phage ZCKP1 like ViPTree analysis Fig. 10B.

Discussion
This study was designed to find lytic bacteriophages against field isolates of APEC ,having suitable characteristics for their putative biocontrol / therapy application against MDR poultry strains of zoonotic potential circulating in capitol territory Pakistan. There were similarities in phylogroup distribution, virulence profiles and resistance patterns in E. coli strains isolated from poultry and human ICU patients. It is established that PCR based phylotyping has 80-95% concordance with MLST analysis, providing an effective, cheaper, and rapid tool to understand the genetic diversity of strains of E. coli 2,3 . Strains from different phylogroups vary in their genome size, gene content, pathogenicity, ecological niche, and life history characteristics 25 . Phylogenetic analyses of E. coli strains from different sources indicate that virulent extraintestinal E. coli strains belong mainly to group B2 and D. In contrast, most of the commensal strains are associated with group A or group B1 4,5 . Our results indicated that phylogroup D was common to both human and poultry isolates. Moreover, most Extraintestinal pathogenic E. coli (ExPEC) causing UTIs in humans belong to phylogroup D 26 . Common virulence determinants of APEC, UPEC and NMEC include traT, iroN, vat, papC, kpsMTII, papEF, ibeA, pic, iss, ompT, cvaC, and fimH etc. 27 . Both poultry and human E coli isolated in our study possess Aggr, ecpA, vat, traT, papC, ompT and fimH as common virulence markers which emphasize their zoonotic potential. One such isolate, QZJM25 from  Table S10B) (C and D) Phylogenetic position of Escherichia phage SKA64. (C) Nucleotide based GDBP tree provides an estimate of phage relationship with close homologs in BLAST n using VICTOR. Analysis identified one Family, one genus, and 14 species clusters respectively, identifying SKA64 as new species.
(D) A proteome tree made by VIP Tree software that uses genome wide similarities calculated by tBLASTx using Similarity score generated by software . Both trees place SKA64 in the same genus. Genomes for comparison were selected according to SG scores of 39 sequences in VIP tree ranging from 1 to 0.07 (supplementary Table S10B). www.nature.com/scientificreports/ phylogroup D having these virulence markers and an extensive drug resistance profile was selected for isolation of E coli specific bacteriophages. As per ICTV recommendations any new bacteriophage can be classified in to one of the existing genera if it has more than 70% nucleotide homology or 40% of its proteins have ≥ 70% homology with members of that genus 28 . Since SKA49 has between 81.5%-85.7% nucleotide homology with members of Carltongylesvirus as well as majority proteins have more than 80% homology to members of this genus it is included in un-classified Carltongylesvirus division. Whereas SKA64 belongs to genus Phapecoctavirus in subfamily Stephanstirmvirinae.
To fulfil ICTV requirements a heat map was generated using VIRIDIC software for genome wide comparison of both phages with their close homologs in respective genera. This heat map identified both SKA49 and SKA64 as new species. SKA49 proteins have high homology with three phages Mangalitsa (27%), ST32 (26%) and GJ1 (15.1%), however only ST32 is physically characterized and for remaining two only genomes have been reported 29,30 . Likewise, for SKA64 only genome sequence of few closely related homologs have been reported but none of them is fully characterized 31,32 .
Several studies have reported characterization and use of bacteriophages against APEC strains of poultry origin. A study performed by Kazibwe et al. isolated and characterized seven lytic bacteriophages against APEC strains isolated from broilers. These phages had moderate host ranges and were not stable at temperatures above 60 °C 19 . Another study conducted by Tang et al., also report similar findings and good potential of lytic phage CE1 to act as surface disinfectant as well as its therapeutic potential to control APEC challenge in vivo in 1-day old broiler chicken 21 .
Utility of any phage in biocontrol / therapy depends on a list of characteristics defined in detail by several publications 22,[33][34][35] First requirement is that phage should be exclusively lytic to control the target bacteria to sub-infectious levels 36 . Both SKA49 and SKA64 are predicted to be lytic as their genomes do not have any ORFs involved in integration and recombination as indicated by PhageLeads software analysis. Both phages were also free of any virulence, toxin, and antimicrobial resistance genes. SKA64 possesses a recombinase family protein www.nature.com/scientificreports/ (ORF101) that is resolvase/ invertase probably for keeping the phage DNA concatemers in single molecules. Same is true for other phages of their respective genera. Lysis potential of both phages cannot be compared with other phages of their respective genera as none of them is tested for their ability to control bacterial growth. However, both SKA49 and SKA64 were able to reduce the bacterial growth keeping it significantly lower than the control for 18 h. They were not tested beyond 18 h. Both phages were tested for bacterial growth reduction at two different MOIs and the difference between their efficacy in controlling bacterial strain was statistically non-significant. Generally higher MOIs are more efficacious than low MOI 37 however, the assay was carried out with slight modification where bacteria and phage were mixed in 1;1 and 1;100 ratio in 1 ml and then diluted to final volume of 40 ml after sufficient phage absorption. This may allow better contact of phage and bacteria to cause efficient infection. Infection of phage to any host is presumed to be controlled by phage to bacterium ratio however MOI does not give account of the total infected bacteria 38 . Similar findings were reported by Liu et al. where MOI value had no significant effect on bacterial infection and appearance of resistance mutations till 80 generations 39 .
Another desired characteristic of phages isolated for biocontrol and therapeutic applications is their stability at normal temperatures for storage. Temperature tolerance analysis indicated that both phages were stable at 37 °C. The same was true for a close homolog of SKA49, Escherichia phage ST32 who was stable between 20 and 37 °C 30 . SKA49 was not tested at 20 °C. However, another member of the same genus, Bp10 was reported to be stable at 30 °C-60 °C temperature 40 . SKA49 was stable at pH 7 and was able to grow at pH5 ad 9 however with a significant drop in titer. Bp10 its close homolog however was stable at all three pH values with no drop in titer. pH stability was not reported for ST32. SKA64 was stable at only pH7 and did not produce any plaques at other pH values. Inability of both phages to stay stable at wide range of pH values may restrict their application in therapeutic applications in vivo to only nasopharyngeal route. Since both phages do not tolerate pH below 5, it may restrict their ability to pass through stomach (pH 2-3.5). However, APEC strains are usually associated with severe respiratory problems which later lead to bacteremia and colibacillosis. Application of phages through nasopharyngeal route/ intra-muscular route may preserve their efficacy. SKA49 and SKA64 may also be employed for biocontrol applications on poultry meat surface . It requires a pH range between 6.5 and 7.3 which is well tolerated by both phages 41 .
SKA49 was effective against 48% of APEC strains tested, whereas SKA64 was effective against only 14% poultry strains tested. One major reason could be that phages were tested against field isolates rather than standardized reference strains and they were not evaluated for presence of any inducible prophages. These prophages if present can make bacterial strain immune to superinfection with another phage 42 . Narrow host range is a major limitation for use of phages in therapy however, such phages can be used in combination with other broad host range phages and antimicrobials to get maximum benefit of their lysis potential 43 . SKA64 tail spike protein is completely unique for members of Phapecoctavirus genus. In a recent study Lone Brøndsted has established that tail spike protein of phages interact with bacterial receptors and subtypes of tail spike protein may shed light on the selected host range of phage. It is likely that SKA64 narrow host range is attributed to this recombination event that led to unique tail spike protein 44 .
The burst size of SKA49 was low (56 pfu/cell), this observation is consistent with the fact that ST32 had a very low burst size at 37 °C but improved when phage grown at 20 °C. Same was true for Bp10 another close homolog of SKA49 whose burst size was 56 virions/cell. Therefore, low virion counts of SKA49 at 37 °C is expected as it www.nature.com/scientificreports/ belongs to cold range phages 45 . Similar comparisons for SKA64 were not possible as none of the close homologs are characterized for their stability, only genomes of few have been reported.
To deploy a phage in therapeutics its genome must be tested for presence of any virulence genes / recombinases and integrases 34 . For this reason, the genome of a therapeutic phage must be characterized. The SKA49 phage genome has 44% and SKA64 has 39% GC content which is the same as reported for other members of their respective genera (Supplementary Tables S10A and S10B). SKA49 has only one tRNA gene (tRNA Arg ) for coding arginine (from 13,796 to 13,887, 92 bp) like ST32 , Flopper 46 and GJ1, however it has no intron in this gene unlike GJ1 47 . Unlike SKA49 , SKA64 possesses 15 tRNA genes. Phages carry tRNA genes to compensate for genomic differences between them and their host. Usually, these tRNA are used for codons highly preferred by phages in contrast to their host [47][48][49] . It indicates that SKA64 has more translational autonomy in comparison to SKA49. These tRNA genes were conserved in the entire Phapecoctavirus genus.
The tail fiber proteins of phages are usually highly mosaic in nature. An excellent example of such mosaic arrangement is ORF 70 , the tail fiber protein of SKA49 where 5' and 3' ends of the ORF seems to be contributed by two different phages (phage Flopper and Mangalitsa respectively). Downstream to the N terminus of this protein a pyocin knob domain has significant protein homology (78%) to phage Flopper only. This pyocin domain is unique to SKA49 and Flopper as no significant BLASTp homology with other phages of the same genus was identified (supplementary Figure S2). These R1 and R2 type of pyocin are powerful bacteriocins used by pseudomonas sp. for eliminating competition 50,51 . Tail spike protein of SKA64 was unique as it has no protein as well as nucleotide homology to any member of Phapecoctavirus genus, rather it resembled (53%QC with ¬ 79% homology) to phages Escherichia phage Pisces and Penshu1, from another genus (Kayfunavirus) of a different family. It agreed with the data presented by Haggard in 1992 indicating that phages can exchange tail fiber genes even though they are not related. It may result in their host range expansion or contraction, and it may allow them to bind alternative receptors on the same host 52 .
SKA49 has two distinct types of endolysins. One of these endolysins is present in the first genome module (ORF22 Fig. 4) and this hydrolase is a D-Ala-D-Gln carboxypeptidase that can lyse D-glutamate and L-alanine residues in glycoprotein portions of bacterial cell wall from within the cell. Such endolysins are part of M15 superfamily of proteins first identified in Listeria monocytogenes (Ply118) 53,54 . Interestingly the N' terminus of this hydrolase till 80AA was conserved in the genus however C' terminus from 80-138AA matches only SA91KD and ST32 phage and for all others it was partially similar. GARD analysis (Genetic Algorithm for Recombination Detection; http:// www. datam onkey. org/ GARD) showed that this hydrolase possesses a recombination hotspot between amino acid 81 and 82 and places endolysin of SKA49, ST32 and SA91KD in one lineage whereas others diverted (supplementary Figures S3A and S3B). similar phenomenon is reported by Oechslin 55 . The second endolysin of SKA49 is a putative SAR endolysin whose ORF clustered together with a pin-holin and a distant spannin (ORF 71,72,& 31, Fig. 4). It belongs to a new class of endolysins called "true lysozymes" 51 . These lysozymes possess an SAR (Signal-Anchor-Release) sequence (as analyzed by HHpred software) that exports this lysozyme to the periplasm in holin independent manner in an inactivated form to prevents premature cell lysis 56 . These enzymes may be able to degrade bacterial cell wall independent of holin by changing protonmotive potential of the outer membrane 57,58 . SKA64 has only one endolysin (ORF 224) that belongs to hydrolase 2 superfamily and is a transglycosylase enzyme 59 .
This detailed characterization of SKA49 and SKA64 will not only add to the existing knowledge about their respective genera it will also pave way for their probable use in therapeutic/ biocontrol applications after in vitro and in vivo evaluations.

Methods
Isolation and confirmation of bacterial strains. Escherichia coli strains used in this study were isolated and characterized in Molecular Virology labs (MVL), CUI to identify endemic Escherichia coli species in poultry showing signs and symptoms of E. coli infection. All methods used in this manuscript have been adapted from established protocols are appropriately cited. For isolation standard protocols as illustrated by 60,61 were used. Briefly, sixty random poultry samples including liver, intestine, caecum, and spleen were collected from different poultry farms in federal area, Pakistan in sterile zip lock bags and stored on ice during transportation.
Thirty patient urine samples with suspected urinary tract infection from ICU patients of three major hospitals in federal area (Rawalpindi and Islamabad), Pakistan were also included in this study. These UTI samples were collected in a previous study at MVL CUI following appropriate guidelines (Helsinki Declaration: https:// www. wma. net/ polic ies-post/ wma-decla ration-of-helsi nki-ethic al-princ iples-for-medic al-resea rch-invol vinghuman-subje cts/) and plated on to MacConkey agar plates on-site in hospital pathology lab by hospital staff after informed patient consent on a pre-approved proforma, and placed in incubator at 37˚C. No human tissue samples were included in the current phage study or sampling was carried out directly from individuals for this study therefore no regulatory approvals were obtained. Poultry samples were processed at MVL CUI. For processing 20 g of each poultry tissue were pooled together in a sterile disposable petri plate and washed with PBS twice. Later this pooled sample was macerated in PBS to fine paste using sterile disposable scalpel blade in a glass Petri plate. Macerated tissue sample was enriched in 10 ml buffered peptone water over night at 37 °C 62 . After 24 h a loop full of enriched medium was streaked on MacConkey's agar (Oxoid; CM0007B) plates and incubated at 37 °C for 18-24 h. Appearance of lactose fermenting pink colonies on MacConkey's agar indicated presence of E. coli. Both poultry and human suspected E .coli strains were further confirmed by biochemical analysis using API 20E kit strips (bioMérieux). Those strains having appropriate biochemical/ numeric profile of API kit as per manufacturer's instructions (Biomerieux) were subjected to molecular confirmation by amplification of 271 bp of E coli housekeeping gene β-glucuronidase (uidA) 63 Table S1). All these strains were used for phage host range identification in further experiments (Table1).
E. coli phylotyping. All E. coli isolates were classified in to appropriate phylogroups using standard quadruplex scheme given by Clermont and colleagues 2012 3 . Briefly the genomic DNA of all E. coli isolates was extracted using PCI (Phenol, Chloroform, Isoamyl alcohol) method 64 and subjected to PCR identification for presence or absence of four genetic markers arpA, chuA, yjaA, TspE4.C2 using oligonucleotide primers given in 3 .
After gel electrophoresis the isolates were either immediately categorized into a phylogroup based on quadruplex genotype or further confirmed using group specific PCR. Phylogroups were assigned based on the schematic presentation as given by Clermont and colleagues in 2012. Phylogroup of all isolates is given in supplementary  Table S1.

Antibiotic sensitivity testing of E. coli isolates. Antibiotic sensitivity of all isolates was tested by
Kirby-Bauer disc diffusion method. phenotype and genotype of all confirmed E. coli isolates was tested against 27 antibiotics commonly employed for treatment of E. coli in humans and poultry. Antibiotic discs with standard recommended concentration were used for sensitivity testing (Oxoid) ( Bacteriophage isolation. One gram tissue sample of intestine, caecum, and stomach from poultry samples previously tested positive for E. coli were pooled and triturated in 10 ml of buffered peptone water and incubated at 37 °C for 24 h without shaking. Next day the samples were centrifuged at 8000 rpm for 10 min (Centrifuge, Hermle, Siemensstr. 25, D-78564 Wehingen, Germany ZK 496) and supernatant was filtered through 0.22 µm syringe filters (CNW Technologies, Düsseldorf, Germany). One milliliter of this supernatant was mixed with 2 ml of QZJM25 overnight culture supplemented with 10 mM CaCl 2 and incubated at 37 °C for 24 h. Next day the mixture was again centrifuged at 8000 rpm for 20 min to remove bacterial cells and supernatant was filtered through 0.22 µm syringe filter. This was crude phage lysate which was stored at 4 °C and later tested by agar overlay spot method as described by 74 . Briefly, 250 µl of QZJM25 overnight culture was mixed with 2.5 ml of soft agar (Lauria-Bertani (LB)-agar 0.5% w/v) and poured on to the LB agar plates. Plates were solidified for 15 min and then 10 µl drops of crude phage lysate were spotted on the plate. Once the drops were dry the plates were inverted and incubated at 37 °C for 24 h. All samples that produced clear zones of phage lysis on bacterial lawn were selected for further purification.
Purification of phages. Usually, the crude phage lysate obtained after phage isolation procedure is a mixture of phages. To purify/ isolate a single phage with good lysis potential the crude phage lysate was subjected to single plaque purification method as describe by 75 77 . To test temperature stability a fixed number of phage particles (2.5 × 10 9 pfu/ ml for SKA49 and 3 × 10 8 pfu/ml for SKA64) were suspended in 1 ml PBS and incubated at 37 °C, 60 °C, and 80 °C for 1 h. After temperature incubation the phage was titrated, and its viability was checked. To determine pH stability the same number of phage particles (2.5 × 10 9 pfu/ml and 3 × 10 6 respectively), were suspended in phosphate buffered saline and adjusted at various pH values (2, 5, 7, 9, and 12) using either 6 M NaOH or 6 M HCl and incubated for 1 h.
One step growth curve. Latent period and burst size of both SKA49 and SKA64 was determined using modification of method given by Kropinski 78 . Briefly, an overnight culture of QZJM25 was diluted (1:100) in 5 ml LB broth and allowed to grow until its OD600 reached log phase (0.3-0.4) at 37 °C. Later this culture was adjusted to 10 6 bacterial cells (CFU/ml) in 1 ml prewarmed LB medium and mixed with respective phage at MOI1 (10 6 pfu) and left without shaking for 30 min to allow phage adsorption. Next the culture was centrifuged (Centrifuge, Helmer, Germany Cat No. ZK 496) at 10,000 rpm for 5 min. Supernatant was removed and stored in a clean tube and titrated to calculate unabsorbed phages, whereas bacterial pellet was resuspended in pre-warmed 30 ml fresh LB in clean flask and allowed to grow. 1 ml aliquots were taken from the culture and immediately centrifuged and supernatant was filtered and saved on ice until titrated. The burst size of phages was calculated using standard formula given in supplementary Table S5.
Host range of phages. Host range of both phages was determined with a modification of agar overlay spot method illustrated by Adams in 1959 79 . All strains showing zones of lysis were further tested by serial two-fold phage dilutions to determine the ability of each strain to allow growth of phages. All strains isolated during this study were used for host range analysis (Table1). Briefly an overnight culture of each strain was grown in 5 ml LB medium. Next day 250 µl of overnight culture of each strain were mixed with 2.5 ml of soft agar (LB containing 0.5% agar) and poured on to LB agar plates already solidified and warmed at 37 °C. Soft agar was allowed to solidify for 20 min and 10 µl drops of phage lysate were placed on them. After drops absorbed, plates were inverted and incubated at 37 °C overnight. Next day plates were monitored for appearance of zones of lysis as indicator of strain sensitivity to phages.

Growth reduction assay of Escherichia coli strain QZJM25.
To test the lysis potential of both phages bacterial growth reduction assay was performed 80 . The overnight culture of QZJM25 was diluted to a final concentration of 1 × 10 3 CFU/ml in 1 ml and mixed with respective phages SKA49 and SKA64 at MOI 1(10 3 pfu) and MOI 100 (10 5 pfu) in triplicates. This mixture was incubated at 37 °C for 30 min to allow efficient phage adsorption. After 30 min this mixture was added to 40 ml LB broth in 100 ml flask for each phage at both MOIs. One flask was left as bacterial control. This experiment was carried out in triplicates. All flasks were incubated at 37 °C for 24 h. Every 2 h, a 1 ml aliquot was removed from each flask and its optical density (600 nm) was measured with spectrophotometer (Cat No. 721-100G, Hinotek, Ningbo, China). A standard curve of QZJM25 CFU/ ml at various optical density values (OD 600 ) was used to determine bacterial colony-forming units (CFU) at each growth point (supplementary Figure S1). This curve was used to determine the CFU/ml of QZJM25 in growth reduction assay. Average CFU/ml obtained from three replicates of assay were plotted in graph and statistical significance of results was determined (Fig. 2). To calculate phage intergenomic similarities with other members in BLASTn, VIRIDIC software 87 was used. This software is recommended by ICTV (International Committee on Taxonomy of Viruses), subcommittee of Bacterial and Archaeal viruses (BAVS) and it applies BAVS traditional algorithms to calculate phage genome similarities. VIRIDIC gives best agreement with traditional algorithm that uses percentage identity of two genomes calculated by BLASTN (http:// rhea. icbm. uni-olden burg. de/ VIRID IC/). Easyfig software (freely available at https:// mjsull. github. io/ Easyfi g/) was used to generate genome comparison figure of phages.

Genome sequencing and phylogenetic comparisons of phages.
To determine suitability of both phages as probable candidates for bacterial control their genomes were tested for presence of lysogenic/temperate lifecycle or recombinase genes. They were also tested for presence of virulence/toxins/antimicrobial resistance markers using PhageLeads software available free at www. phage leads. dk 88 .
The phylogenetic analysis was carried out by the VICTOR web service (https:// victor. dsmz. de), a method for the genome-based phylogeny and classification of prokaryotic viruses 89 . All pairwise comparisons of the nucleotide sequences were conducted using the Genome-BLAST Distance Phylogeny (GBDP) method 90 under settings recommended for prokaryotic viruses. The resulting intergenomic distances were used to infer a balanced minimum evolution tree via FASTME including SPR postprocessing 91 using the optimal formulas D0. Tree was rooted at the midpoint 92 and visualized with ggtree 93 . Taxon boundaries at the species, genus and family www.nature.com/scientificreports/ level were estimated with the OPTSIL program (https:// cran.r-proje ct. org/ web/ packa ges/ optpa rt/ optpa rt. pdf), using the recommended clustering thresholds and an F value (fraction of links required for cluster fusion) of 0.5 90 . In addition to this a phylogenetic tree based on proteome analysis of both phages was constructed using ViPTree software 94 freely available at https:// www. genome. jp/ viptr ee/. To strengthen the phylogenetic relationship of both SKA49 and SKA64 and verify its classification CoreGenes 5.0 software was used to identify common genes shared with their close homologs 95 . The software is freely available at https:// coreg enes. ngrok. io/. The amino acid sequence of two orthologs, Major capsid protein (SKA49) and Tail fiber protein (SKA64) identified by CoreGenes analysis were used to construct maximum likelihood tree (MLT). Briefly Major Capsid protein (MCP) and Tail Fiber protein (TFP) amino acid sequences were aligned together in Geneious (v.8.1.9) 96 . The alignments in Phylip format were imported in IQtree 97 online version 98 available at http:// iqtree. cibiv. univie. ac. at/ to reconstruct a Maximum Likelihood tree based on the best fit model as implemented in the IQTree, using 100 bootstrap value. IQTree implements Model Finder 99 to calculate best fit model on the data. The trees in Newick format were refined in the online TreeDyn tool 100 available at http:// www. phylo geny. fr/ one_ task. cgi? task_type = TreeDyn and downloaded in pdf format.
Statistical analysis. The data analyses were carried out using XLSTAT free 2021 (Addinsoft, Paris-France) and OriginPro 2022 (OriginLab Corporation, Northampton, MA, USA). Depending upon the type of study, the significant difference between variables was assessed using either a student's t-test or an ANOVA with a P value of 0.05 or less (p ≤ 0.05) was considered statically significant.
Ethics approval. E coli isolates from urine samples included in this study for comparisons were collected in a previous project approved by institutional ethical review board with letter number CUI/Bio/ERB/6-18/02. All samples were collected after informed patient consent on a pre-approved proforma in native language. Samples were collected following Helsinki guidelines https:// www. wma. net/ polic ies-post/ wma-decla ration-of-helsi nkiethic al-princ iples-for-medic al-resea rch-invol ving-human-subje cts/ .

Funding
The project for phage isolation was carried out by IFS (International Foundation of Science) funding vide grant number I-3-B-6115-1.