In Vitro and In Vivo Assessments of Two Newly Isolated Bacteriophages against an ST13 Urinary Tract Infection Klebsiella pneumoniae

Antibiotic resistance represents a major public health concern requiring new alternatives including phage therapy. Klebsiella pneumoniae belongs to the ESKAPE bacteria and can cause urinary tract infections (UTIs). The aims of this study were to isolate and characterize new bacteriophages against a K. pneumoniae strain isolated from UTIs and to assess their efficacy in vitro and in vivo in a Galleria (G.) mellonella larvae model. For this purpose, two bacteriophages were newly isolated against an ST13 K. pneumoniae strain isolated from a UTI and identified as K3 capsular types by wzi gene PCR. Genomic analysis showed that these bacteriophages, named vB_KpnP_K3-ULINTkp1 and vB_KpnP_K3-ULINTkp2, belong to the Drulisvirus genus. Bacteriophage vB_KpnP_K3-ULINTkp1 had the narrowest host spectrum (targeting only K3), while vB_KpnP_K3-ULINTkp2 also infected other Klebsiella types. Short adsorption times and latent periods were observed for both bacteriophages. In vivo experiments showed their ability to replicate in G. mellonella larvae and to decrease host bacterial titers. Moreover, both bacteriophages improved the survival of the infected larvae. In conclusion, these two bacteriophages had different in vitro properties and showed in vivo efficacy in a G. mellonella model with a better efficiency for vB_KpnP_K3-ULINTkp2.


Introduction
Klebsiella pneumoniae can roughly be classified into two pathotypes: the opportunistic ones, which are often multidrug-resistant (mdrKp), and the hypervirulent ones (hvKp). Today, drug-resistant properties coexist with hypervirulent ones [1], and this bacterium represents a huge public health concern. Indeed, "33,000 deaths per year in the EU and €1.5 billion per year in healthcare costs and productivity losses" [2] are figures related to antimicrobial resistance and K. pneumoniae, a member of the ESKAPE pathogens (Enterococcus faecium, Staphylococcus aureus, Acinetobacter baumannii, Klebsiella pneumoniae, Pseudomonas aeruginosa, and the Enterobacter species) [3], is on the WHO's "critical priority" list of

Bacterial Strain
The bacterial strain selected for the bacteriophages' isolation, K. pneumoniae QAMH 130326/0185, was previously isolated from a urine sample of a Belgian diabetic patient with a UTI. Its capsular type (K-type, K) was identified by wzi gene PCR [5], and the phenotypic antibiotic susceptibility profile of this strain was assessed by disk diffusion assay using Mueller-Hinton plates (Becton Dickinson, Erembodegem, Belgium), Axonlab disks (Belgium), and the SIRscan micro software (Axonlab, Machelen, Belgium). The determination of the colistin minimum inhibitory concentration (MIC) was performed by broth dilution assay in FRCOL 96-well microtiter plates, according to the manufacturer's instructions (Sensititre, Fisher Scientific, Merelbeke, Belgium) [26]. The results were interpreted using EUCAST (European Committee on Antimicrobial Susceptibility Testing) clinical breakpoints (2021 edition) and CASFM VET (Comité de l'Antibiograme de la Société Française de Microbiologie) standards (2019 edition).

Bacteriophage Isolation
In order to isolate new bacteriophages, 5 mL of the supernatants of centrifuged (10 min at 9500× g) and ultrafiltered (0.2 µm) wastewaters were brought into contact with the K. pneumoniae QAMH 130326/0185 (100 µL at optical density (OD) (600 nm) between 0.2 and 0.3) in 5 mL of twice concentrated Luria-Bertani (LB) broth (Sigma-Aldrich, Saint-Louis, MO, USA) enriched with MgSO 4 (1 mM) and CaCl 2 (1 mM). This solution, as well as a sterility control of the wastewater and a bacterial growth control, were incubated at 37 • C for 4 h until bacterial lysis. The lysate and the bacterial growth broths were centrifuged (9500× g for 10 min) and filtered (0.2 µm) and were spread (4 µL) on enriched LB agar with a K. pneumoniae QAMH 130326/0185 bacterial overlay technique. After incubation at 37 • C for 4 h, one individual plaque was selected and purified four times following the same procedure. Transmission electronic microscopy (TEM) pictures were obtained on purified bacteriophages using a Tecnai Spirit microscope (FEI, Eindhoven, The Netherlands) operating at 120 kV at the electron microscopy unit of Sciensano (Bruxelles, Belgium). The bacteriophages were purified by CsCl density gradient (layers of 1.33, 1.45, 1.50, and 1.70 g/cm 3 ) ultracentrifugation (133,900× g; 3 h; 4 • C), dialysis using Slide-A-Lyzer dialysis cassettes G2 (Thermo Fisher Scientific Inc., Merelbeke, Belgium). Four µL of both bacteriophages were plated on a K. pneumoniae QAMH130326/0185 overlay to analyse the plaque of lysis morphologies.

Bacteriophages Genome Sequencing
To analyse the genomes of these bacteriophages, genomic DNA was extracted from CsCl purified bacteriophages using the DNeasy Blood & Tissue kit (Qiagen, Crawley, UK). The sequencing libraries were obtained using the Illumina Nextera XT DNA sample preparation kit and sequenced using a MiSeq Illumina Next instrument (Illumina, San Diego, CA, USA). The bacteriophage genomes were assembled into single contigs using SPAdes Assembler V3.10 [27]. The web server RAST was used for genome annotation with coding DNA sequences (CDS) outputs. These results were then compared with the web server HHpred (https://toolkit.tuebingen.mpg.de/tools/hhpred, accessed on 1 February 2022) [32]. Homologous related bacteriophages were found in GenBank using BLASTn analysis (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE_TYPE=BlastSearch, accessed on 1 February 2022), and closely related bacteriophage genomes were illustrated with Easyfig [33]. The automated prediction of the cycle of both bacteriophages was assessed with PhageAI [34].
After the investigation of the normality of each distribution (corresponding to the titration data for each phage and for both incubation condition) (by a histogram, a quantilequantile plot (QQ-plot), a boxplot, and a Shapiro-Wilk test), a one-way analysis of variance (ANOVA) (with multiple comparisons) was assessed to highlight if the concentration obtained after 1 h incubation was significant different from the original. All statistical analyses were performed with R (p-value ≤ 0.05) [35].

Host Range and Efficiency of Plating
The host range of both bacteriophages was assessed using 31 different Klebsiella and 8 UPEC E. coli strains (Table S1). Each bacteriophage was spotted (4 µL) in duplicate on the different bacterial overlays (OD: 0.15-0.25) and incubated at 37 • C until the appearance of lysis plaques. The results were divided in three classes: no lysis, confluent lysis or individual lysis plaques. Then, if there was lysis (confluent lysis or just individual plaques), the efficiency of plating (EOP, the ratio between the concentration of the tested strain over the concentration of the propagation strain) was determined (Table S1). To that end, each bacteriophage was tenfold serially diluted in PBS (in triplicate), drops (2 µL) were plated on enriched LB agar with the different strains overlayers, and Petri dishes were then incubated for 4 h at 37 • C.

Adsorption Times and Low MOI Kinetic Curves
The adsorption times and the low multiplicity of infection (MOI) kinetic curves (MOI of 0.1) were determined in biological triplicate. For the adsorption time, the bacteriophage was mixed with the host bacterium (QAMH 130326/0185) at MOI of 0.1 and incubated for 10 min at 37 • C. Every 2 min, 100 µL of the solution was sampled and 10 times diluted in PBS to reduce the interaction of the bacteriophage with the bacteria. After 10 min, all samples were filtrated (0.2 µm), and the unadsorbed bacteriophages were titrated to measure the adsorption time [36]. The adsorption rate constants were calculated according to the following mathematical formula: where k is the adsorption rate constant, in mL/min; B is the concentration of bacterial cells; and t is the time interval in which the titre falls from P0 (the original concentration of the bacteriophage) to P (the final concentration of the bacteriophage) [37]. For the low MOI kinetic curves, a mix of bacteriophage and bacteria similar to that prepared for the adsorption times was performed in biological triplicate, incubated at 37 • C till the bacteriophage's adsorption, and then centrifugated (13,000× g, 1 min). The pellet was resuspended in 10 mL of LB Lennox and incubated at 37 • C with shaking over 100 min. Every 5 min, 100 µL were sampled, 10 times diluted in PBS, filtrated (0.2 µm), and then titrated to highlight the latent periods and the steady state times.

In Vitro Bacteriophage Activity
A total of 100 µL of K. pneumoniae QAMH130326/0185 (at 10 6 colony formCFU/mL) and different concentrations of each bacteriophage (1000 µL, 100 µL, and 10 µL at 10 8 PFU/mL for MOIs 1000, 100, and 10) were mixed in enriched LB Lennox broth (5 mL) in biological triplicates. Three control tubes were added to the experiment: a growth control, a sterility control, and a bacteriophage control. All tubes were incubated at 37 • C with shaking, and optical densities (OD 600) were measured every hour for 12 h.

Survival Rate in Galleria mellonella Larvae
The optimal bacterial inoculation dose which can kill 80% of the larvae within 4 days was determined in one preliminary experiment (data not shown). Then, the assessment of the survival rates for infected larvae treated with the bacteriophages was performed in a single experiment for MOIs of 10 and 100. A total of 150 larvae were distributed in 5 groups (Table 1), and each larva was inoculated using an automatic injector (Cole Parmer, Vernon Hills, IL, USA) in the last left proleg for the first injection followed by a second injection 1 h later in the last right proleg. Injections of 10 µL were performed using BD Plastipak™ 1 mL sterile syringes (Becton-Dickinson, Franklin Lakes, NJ, USA) and sterile 30-gauge needles (Terumo corporation, Tokyo, Japan). The larvae were incubated at 37 • C, and their survival was evaluated every 24 h. Kaplan-Meier survival curves were generated to assess the survival of the different groups using Rcommander [27]. Log-rank tests were performed to highlight any significant difference in survival rates between the groups (p-value ≤ 0.05). Then, if the MOI of 100 was not sufficient to improve the survival rates of the insects, another single experiment with an MOI of 1000 was performed (Table 1).

Bacterial Load and Bacteriophage Replication in Galleria mellonella Larvae
The second main in vivo experiment aimed to assess the bacterial and bacteriophage titre evolutions [36,38]. For this single experiment, 120 larvae were divided into 6 groups and inoculated following the same protocol (using an MOI of 100). Bacterial and bacteriophage titrations were performed after 24 h and 72 h in technical triplicate, with larvae ground in a homogenizer (Stomacher) in groups of 10 individuals. The mix (200 µL) was diluted in PBS and weighed. For the bacteriophage titre, half of the sample was filtered using a 0.2 µm sterile syringe filter (ref.514-0073, VWR, Leicestershire, UK), 10-fold serially diluted in PBS, plated (2 µL) on LB agar with a QAMH 130326/0185 overlay and incubated at 37 • C 4 h. For the bacterial titration, the diluted samples were plated on SCAi agar (Simmons Citrate Agar, Inositol 1% and ampicillin 2 µg/mL, Sigma-Aldrich, Saint-Louis, MO, USA) and incubated at room temperature for 24 h.
After the investigation of the normality of each distribution (corresponding to the bacterial and phage titration data for each phage) (by a histogram, a quantile-quantile plot (QQ-plot), a boxplot, and a Shapiro-Wilk test), a two-way non-parametric ANOVA (with multiple comparisons) was assessed to highlight if there were significant differences between groups and during all the experiments. All the statistical analyses were performed with R (p-value ≤ 0.05) [35].
The genomic analysis assigned this 5,628,480 bp length bacteria to the sequence type (ST) ST13. Two virulence factors were found: the genotoxin colibactin clb-3 and the yersinibactin ybt-17, both included in integrative and conjugative element (ICE, ICE-Kp10. Regarding the antibiotic resistance elements, 3 acquired (bla OXA-1 , sul-1 and aadA1) and 3 intrinsic (fosA, oqxAB and blaSHV-1) genes were found. At the protein level, proteins FimA, B, D, F, G, and H were predicted.
Sequencing was submitted as NCBI BioProject PRJNA821689.

Bacteriophage Isolation
Two bacteriophages (vB_KpnP_K3-ULINTkp1 and vB_KpnP_K3-ULINTkp2) were isolated from wastewaters collected in Belgium in 2020 (Brussel (North Station) for vB_KpnP_K3-ULINTkp1 and Liège (Oupeye) for vB_KpnP_K3-ULINTkp2). The TEM pictures showed a Podoviridae morphology with a non-enveloped icosahedral head, a diameter of approximately 55 nm, and a very short tail for both bacteriophages (Figure 1). The lysis plaques showed halo zones and diameters of around 3 mm and 1 mm respectively for vB_KpnP_K3-ULINTkp1 and vB_KpnP_K3-ULINTkp2 ( Figure S1).
Sequencing was submitted as NCBI BioProject PRJNA821689.

Bacteriophage Isolation
Two bacteriophages (vB_KpnP_K3-ULINTkp1 and vB_KpnP_K3-ULINTkp2) were isolated from wastewaters collected in Belgium in 2020 (Brussel (North Station) for vB_KpnP_K3-ULINTkp1 and Liège (Oupeye) for vB_KpnP_K3-ULINTkp2). The TEM pictures showed a Podoviridae morphology with a non-enveloped icosahedral head, a diameter of approximately 55 nm, and a very short tail for both bacteriophages (Figure 1). The lysis plaques showed halo zones and diameters of around 3 mm and 1 mm respectively for vB_KpnP_K3-ULINTkp1 and vB_KpnP_K3-ULINTkp2 ( Figure S1).
The sequences of the phages were submitted on the NCBI BioProject PRJNA821689.
The sequences of the phages were submitted on the NCBI BioProject PRJNA821689.

Temperatures and pH Stability
Both bacteriophages showed stable lytic activity over the pH range of 4-12 (Figure 3), and both bacteriophages were thermostable until at least 60 • C (Figure 4).

Temperatures and pH Stability
Both bacteriophages showed stable lytic activity over the pH range of 4-12 ( Figure  3), and both bacteriophages were thermostable until at least 60 °C (Figure 4).

Temperatures and pH Stability
Both bacteriophages showed stable lytic activity over the pH range of 4-12 ( Figure  3), and both bacteriophages were thermostable until at least 60 °C (Figure 4).

In Vitro Bacteriophage Activity
The in vitro efficacy assessment of the bacteriophages showed a lytic activity for both bacteriophages (irrespective of MOI used) ( Figure S2).

In Vitro Bacteriophage Activity
The in vitro efficacy assessment of the bacteriophages showed a lytic activity for both bacteriophages (irrespective of MOI used) ( Figure S2).

Survival Rate in Galleria mellonella Larvae
Compared with the infected but not treated group, the infected larvae treated with vB_KpnP_K3-ULINTkp1 showed a better survival rate only with an MOI of 1000 (53% vs. 27%), in contrast with vB_KpnP_K3-ULINTkp2, which already improved the survival at a MOI of 10 (67% vs. 23%) (Figure 7).

In Vitro Bacteriophage Activity
The in vitro efficacy assessment of the bacteriophages showed a lytic activity for both bacteriophages (irrespective of MOI used) ( Figure S2).

Survival Rate in Galleria mellonella Larvae
Compared with the infected but not treated group, the infected larvae treated with vB_KpnP_K3-ULINTkp1 showed a better survival rate only with an MOI of 1000 (53% vs. 27%), in contrast with vB_KpnP_K3-ULINTkp2, which already improved the survival at a MOI of 10 (67% vs. 23%) (Figure 7).

Discussion
The two bacteriophages isolated and characterized in this study showed pred virulent cycles, a productive infection with short adsorptions and latent periods d the low MOI kinetic curve, and as such, they could represent an interesting alternat antibiotics. Clear lysis plaques with a halo zone suggest that they possess an exopo charide-degrading enzyme, highlighted in silico [39]. They were stable over large pH temperature ranges, in accordance with activity in human urine, which exhibits ranging from 5.4 to 7.2. The presence of endolysins has been observed in silico, thes drolases digesting the rigid bacterial peptidoglycan wall leading to bacterial lysi death. With a fast bacterial lysis, with a high specificity, with the possibility of mole engineering to change the host range or allow to cross the outer membrane, with ac

Discussion
The two bacteriophages isolated and characterized in this study showed predicted virulent cycles, a productive infection with short adsorptions and latent periods during the low MOI kinetic curve, and as such, they could represent an interesting alternative to antibiotics. Clear lysis plaques with a halo zone suggest that they possess an exopolysaccharide-degrading enzyme, highlighted in silico [39]. They were stable over large pH and temperature ranges, in accordance with activity in human urine, which exhibits a pH ranging from 5.4 to 7.2. The presence of endolysins has been observed in silico, these hydrolases digesting the rigid bacterial peptidoglycan wall leading to bacterial lysis and death. With a fast bacterial lysis, with a high specificity, with the possibility of molecular engineering to change the host range or allow to cross the outer membrane, with activity against biofilms, and with a low risk of resistance emergence, these so-called enzybiotics seem promising [40]. In addition, putative endonucleases were found. These enzymes are believed to reduce the horizontal gene transfer by transformation by avoiding the release of intact plasmid in the environment [41].
The K-type 3 (K3) clusters are observed only in K. pneumoniae species. Their capsules are rich in mannose residues [42] used for interactions with mannose receptors that are particularly present on macrophages' surfaces [43]. K3 includes two subspecies: Klebsiella pneumoniae subsp. pneumoniae and Klebsiella pneumoniae subsp. rhinoscleromatis. The subspecies pneumoniae could be isolated from fresh urine or from stool of diarrheic patients [44]. The subspecies rhinoscleromatis is the agent of nasal and upper respiratory chronic infection, rhinoscleroma. The latter can be genetically differentiated from the other by MLST, by the absence of kfu (an iron uptake marker), and by the presence of rmpA (regulator of mucoid phenotype A) [45]. With the type 1 fimbrial adhesin, K. pneumoniae has the capacity to colonize the urinary tract (UT) as the putative protein FimA is responsible for bladder cell invasion and biofilm development [46,47], FimH can link the mannose-binding receptor of the UT [47], and FimD, F, and G are minor subunits also involved in UT invasion [48]. Although some ST13 K. pneumoniae are multidrug-resistant bacteria [49,50], this was not the case for the strain QAMH130326/0185 even when it was phenotypically resistant to fosfomycin, a first-line urinary antibiotic.
Klebsiella bacteriophage specificity for the capsular type is well-known [25,51]. This specificity could be linked to a specific interaction of the bacteriophage's depolymerase, highlighted in silico for both bacteriophages in this study, with the bacterial capsule [52][53][54], even if a broad host range of K. pneumoniae bacteriophages have already been described [24,55]. In this study, vB_KpnP_K3-ULINTkp1 seemed more specific than vB_KpnP_K3-ULINTkp2 and could replicate within two of the tested strains (QAMH 130326/0185 and SB5904) that both belonged to K-type 3. Conversely, vB_KpnP_K3-ULINTkp2 could replicate in four other K-type strains but with a low production EOP (between ≤0.1 and ≥0.001). The host range and the EOP showed that both bacteriophages were unable to replicate in some strains but still produced lytic spots at high concentrations. This phenomenon, reflecting a putative abortive infection, could represent a non-productive bacteriophage infection which kills the bacteria [56].
The bacteriophages phiKpS2, vB_KpnP_NER40, and vB_KpnP_SU503 were genomically related to these newly isolated bacteriophages. The first has been isolated with the K. pneumoniae DSM2026 mutant (K. pneumoniae S2) in an abnormal propanediol fermentation broth [57]. The second had been isolated in the Chermyankan river with a K2 K. pneumoniae (Kp40) and showed a narrow spectrum to this capsular type [58]. The third had been isolated using the K. pneumoniae strain 07RAFM-KPN-503, and could replicate, throughout the host range experiment, in another K. pneumoniae strain, 07RAFM-KPN-510 [59,60].
The behaviours of vB_KpnP_K3-ULINTkp1 and vB_KpnP_K3-ULINTkp2 differ in vivo. Indeed, the infected G. mellonella larvae treated with vB_KpnP_K3-ULINTkp1 showed a higher survival rate only at MOI of 1000 with 53% of survival instead of 27% in the infected and nontreated group. The treatment with the vB_KpnP_K3-ULINTkp2 bacteriophage seems to be more efficient, with 77% survival at MOI of 100 and 67% at MOI of 10 instead of 23% for the infected larvae. The bacterial titration experiments showed that both bacteriophages seemed to decrease the bacterial load compared with the infected group. However, the efficacy of vB_KpnP_K3-ULINTkp2 was better than that of vB_KpnP_K3-ULINTkp1 with a lower bacterial load at 24 h and 72 h. Both bacteriophages seemed to replicate in the larvae, with higher concentrations obtained in the treated group.

Conclusions
Two newly isolated Klebsiella bacteriophages demonstrated their efficacy in vitro and in vivo by increasing the survival of G. mellonella larvae. Even if this did not result in a complete elimination of the inoculated bacteria, these results suggest that the two studied bacteriophages could efficiently prevent a K. pneumoniae urinary tract infection induced by this ST13 KL3 K3 strain. Although these bacteriophages are genomically closely related, vB_KpnP_K3-ULINTkp2 showed better efficacy in the in vivo G. mellonella model and a broader host spectrum than vB_KpnP_K3-ULINTkp1. Both conditions being interesting, a narrow spectrum avoids a microbiota disruption in complex microbial systems such as the intestine, but a broad spectrum increases the success of treatment in less complex microbial systems such as the urinary tract, where Klebsiella is not usually found in healthy people.