New Bacteriophages Members of the Ackermannviridae Family Specific for Klebsiella pneumoniae ST258

Abstract

Background:

Carbapenem-resistant Klebsiella pneumoniae, particularly isolates classified as sequence-type 258 (ST258), are multidrug-resistant strains that are strongly associated with poor-prognosis nosocomial infections, as current therapeutic options are limited and ineffective. In recent years, phage therapy has emerged as a promising treatment option for these scenarios.

Methodology and Results:

We report the isolation and characterization of three new phages against Klebsiella pneumoniae ST258 strains recovered from Machángara river wastewater. These new members of the Ackermannviridae family showed stability over a wide temperature and pH range and burst sizes ranging from 6 to 44 plaque-forming units per bacteria. Their genomes were about 157 kilobases, with an average guanine-cytosine content of 46.4% and showed presence of several transfer RNAs, which also allowed us to predict in silico a lytic replicative cycle due to the presence of endolysins and lysozymes.

Conclusion:

Three lytic phages of Ackermannviridae family were recovered against Klebsiella pneumoniae ST258 strains from sewage; however, further characterization is needed for future consideration as therapeutic alternatives.

Introduction

The overuse of antibiotics has led to rapid spread of multidrug-resistant (MDR) Klebsiella pneumoniae virulent lineages, which are associated with high mortality rates especially in infections by isolates belonging to clonal complex 258, where isolates with sequence-type 258 (ST258) are considered as an “high-risk” clonal group causing a large proportion of infections, associated with worse prognosis.^1–6 Among the alternative treatments proposed to address this public health crisis, bacteriophage (phage) therapy has demonstrated efficacy in experimental animal models and shows promise in clinical cases involving MDR pathogen infections, including ST258 strains.^7–10

Moreover, the most common sources of isolation are relatively easy to access, such as sewage and hospital wastewater.^11–14 For example, a recent study published by Hesse et al.⁸ evaluated the efficacy of early treatment with two phages, termed P1 and P2, in reducing bacterial load in infected mice with refractory bacteremia. The results of the study showed that early treatment with both phages, either individually or in combination, resulted in a significant reduction in bacterial load compared with placebo-treated animals. In addition, an improvement in survival was observed in mice treated with the phages compared with controls.⁸

Despite some promising results with other high-risk clonal lineages such as ST11, ST15, and ST16, currently available evidence remains insufficient for consideration beyond experimental treatments. Considering the current challenge posed by MDR lineages of K. pneumoniae, this study aims to address the knowledge gap by in vitro characterization of phages recovered from wastewater and evaluation of their lytic potential on ST258 strains for future development of effective therapeutic strategies.

Materials and Methods

Host strains, sample collection and isolation, enrichment, titration, and purification

Two clinical strains of Klebsiella pneumoniae carbapenemase (KPC)-producing Klebsiella pneumoniae ST258 (14-765 and 14-751) donated by the Instituto Nacional de Investigación en Salud Pública–Leopoldo Izquieta Pérez (INSPI-LIP) from Ecuador were used as host bacteria. The results of antimicrobial susceptibility testing are in Supplementary Table S1. To obtain the exponential phase, 500 μL of a saturated broth (obtained from one or two isolated MacConkey agar colonies mixed in 4.5 mL of Tryptic Soy Broth [TSB] at 37°C and constant shaking at 9 × g overnight) was used beforehand. Next, 500 μL of the aforementioned mixture was added to 4.5 mL of TSB and incubated for 4 h at 180 rpm. A standard inoculum of 0.5 McFarland (1.5 × 10⁸ CFU/mL) was used for all assays.

Samples of wastewater were collected from two different points of Machángara river, Quito-Ecuador. Sample named K0601 was collected from point one, coordinates: 0°12′34.7″ S 78°28′36.6″ W and Sample T0701 from collection point two, coordinates: 0°16′37″ S 78°31′14″ W, all samples were stored at 4°C for 24 h for sedimentation.

Bacteriophage isolation was performed according to the protocol described by Kropinski et al.¹⁵ and Clokie and Kropinski¹⁶ with certain modifications. In brief, 1 mL of the host strain in exponential phase culture, 49 mL of doubly concentrated TSB supplemented with 10 mM Ca²⁺ and Mg²⁺, and 50 mL of supernatant, previously filtered with filter paper (10 μm pore size filter) to separate the suspended solids present in the wastewater samples, were mixed. The mixture was incubated for 24 h at 37°C. The mixture was centrifuged at 1475 × g for 10 min to remove cells, and the supernatant was then filtered through a 0.22 μm pore size filter and stored at 4°C.

Phage presence was verified using the spot test, for which 50 μL of the filtered supernatant was dropped onto the host strain inoculated by extension on Tryptic Soy Agar supplemented with 10 mM Ca²⁺ (TSAd10) and incubated overnight.

For phage isolation, the double layer agar (DLA) technique was used using nutrient agar supplemented with 10 mM Ca²⁺. Serial dilutions and the DLA technique were performed with the previously filtered supernatant. In brief, 100 μL of exponential phase culture, 100 μL of dilution, and 5 mL of agar (top layer) were mixed and poured over the base layer and incubated. Procedure was repeated with all serial dilutions. Phage concentration was expressed in plaque-forming units per milliliter (PFU/mL). We selected potential phages with concentrations over 1 × 10¹⁰ PFU/mL.

For purification, a lysis plaque (10^5–10⁸ PFU/mL) was taken, resuspended in 900 μL of buffer (MgSO₄∙7H₂O 8 mM, Tris-HCl 50 mM, pH 8.4) and shaken vigorously to dislodge the phage from the agar. The mixture was inoculated by extension on TSAd10 and left to stand for 30 min. Later, 500 μL of exponential phase culture was mixed with 5 mL of TSAd10 on top of the aforementioned agar and incubated for 20 h. The best defined and spaced plaques formed after 20 h were picked and stored at 4°C in 1 mL of buffer and 10 μL of chloroform for preservation

In vitro phage activity and stability

Phage characterization was performed by the method described by Manohar et al.¹⁷ Thus, 100 μL of exponential phase culture host bacteria and 100 μL of purified phage was mixed and incubated for 25 min. After centrifugation, the pellet was homogenized using 10 mL of TSB, the DLA technique was performed to obtain the number of phages. Burst size was calculated dividing the final number of free phages by the initial number of phages. For pH stability assay we added 100 μL of purified phage in 900 μL of saline solution adjusted to different pH and incubated for 4 h, whereas for thermostability assay a volume of 100 μL of purified phage was maintained at different temperatures for 1 h.

All assays were performed in triplicate and the DLA technique was applied. Adsorption rate was performed according to reference.¹⁸ In brief, 100 μL of the exponential phase host strain was mixed with 100 μL of purified phage, allowed to stand for 5 min at room temperature, 100 μL of this mixture was added in 9.9 mL of TSB and incubated for 1 h, every 10 min the DLA method was performed. Adsorption constant was calculated with the following equation:¹⁵ $K=\frac{2.3}{B\ast t}log\frac{P_o}{P}$, where B is the initial concentration of bacteria, Po the initial titer, P the final titer, and t the time interval between Po and P. The constant was expressed in mL/min.

Genomic DNA extraction

The phage was suspended in buffer, concentrated by centrifugation and filtered through a 0.22 μm pore-size filter. DNA extraction was performed using Wizard Promega protocol and was modified by adding proteinase K and DNase I. 1 mL of concentrated phage was mixed with 10 μL of DNase I, 1 μL of RNase A, 500 μL of lysis solution and 4 μL of proteinase K, shaken and incubated at the following temperatures: 55°C (60 min), 65°C (15 min), then 320 μL isopropanol was added. Six hundred fifty microliters of the mixture was transferred to a spin column and centrifuged (1 min at 6000 g). Wash solution (90% ethanol) was added 400 μL and centrifuged again. Finally, we added 75 μL of rehydration buffer to the column, centrifuged again and obtained phage genomic DNA, which was stored at −20°C.

Complete genome sequencing and analysis

Sequencing was performed by Biosequence (Ecuador) in the Illumina MiSeq system. Raw sequencing data were deposited in the Sequence Reads Archive: BioProject: PRJNA815380, BioSamples: SAMN26549669 (K751), SAMN26549667 (T751), and SAMN26549668 (T765). Reads obtained were assembled de novo using Unicycler¹⁹ in PATRIC²⁰. Quality and integrity of the viral metagenome-assembled genomes were evaluated with CheckV²¹ and subsequently annotated with Prokka²² in Galaxy.²³

Genomes were aligned and rearranged with MAUVE.²⁴ Prokaryotic virulence, toxin encoding, and antimicrobial resistance genes were searched for using Bakta.²⁵ Our genomes recovered were deposited in GenBank: ON202820 (K751), ON323462 (T751), and ON399185 (T765). Taxonomic assignment was performed by KRAKEN2²⁶ and average nucleotide identity (ANI) was calculated to support it.²⁷ The phylogenomic tree of the whole-genome sequences were generated by VICTOR.²⁸ The replicative cycle was predicted in silico with BACPHLIP²⁹ and PhageAI.³⁰

Results

Phenotypic profile

Phages presence was evidenced by formation of lysis zones and plaques of variable size (ø 1.5–2.0 mm) with a translucent center and a surrounding halo (Supplementary Figs. S1 and S2). Three possible phages were isolated and named according to the recommendations of the International Committee on Taxonomy of Viruses:³¹ (Table 1) Klebsiella_virus_K751 (K751), Klebsiella_virus_T751 (T751) and Klebsiella_virus_T765 (T765). T751 and T765 were inactivated at temperatures >60°C, whereas K751 was completely inactivated at 80°C (Supplementary Fig. S3).

Table 1.

Characteristics of Phage Isolated Against Carbapenem-Resistant Klebsiella pneumoniae ST258 Strains

	K751	T765	T751
Host strain	14–751	14–765	14–751
Maximum concentration (PFU/mL)	1.0 × 1011	9.0 × 1011	2.0 × 1010
Phenotypic profile
Temperature (°C)	−16 to 70	−16 to 60	−16 to 60
pH	3 to 11	3 to 11	4 to 11
Adsorption constant, K (mL/min)	2.1 × 10−10	5.1 × 10−10	1.1 × 10−10
Adsorption at 10 min (%)	28.4	24.4	12.1
Burst size (PFU/bacterial)	6	10	44
Latent period (minutes)	35	45	35
Taxonomy
Family	Ackermannviridae	Ackermannviridae	Ackermannviridae
Genus	Taipeivirus	Taipeivirus	Taipeivirus
Genomic profile
Size (bp)	157.100	157.384	157.384
Depth	125.1 ×	219.2 ×	274.7 ×
Content GC (%)	46.5	46.4	46.4
Integrity (%)	94.9	94.9	94.8
ORF	197	195	196
CDS	204	205	204
Hypothetical proteins	188	188	190
Functional proteins	23	24	21
tRNA	5	6	6
sRNA	1	1	1
In silico replicative cycle (%)
PhageAI (lytic)	93.45	93.05	93.05
BachliB (lytic)	97.02	97.02	97.02

CDS, coding DNA sequences; GC, guanine-cytosine content; ORFs, open reading frames; PFU, plaque-forming units; sRNA, small RNA; tRNA, transfer RNA.

Regarding pH, the candidates survive in an environment of pH 4–11, and are completely inactivated at pH 12. K751 and T765 tolerated pH 3 (Supplementary Fig. S4). K751 and T751 reported a latency period of ∼35 min, whereas T765 was 45 min. Burst sizes (PFU/bacterial) were 6, 44, and 10, respectively (Fig. 1). In addition, at 10 min, it was estimated that 28.4% of K751, 12.1% of T751 and 24.4% of T765 viral particles were adsorbed on the host cell (Table 1).

FIG. 1.

Kinetics of Klebsiella_virus_K751, Klebsiella_virus_T751, and Klebsiella_virus_T765. The relationship between concentration and incubation time with their host cell is shown. Results are based on three replicates. PFU, plaque-forming units.

Genomic profile

Each recovered genome was characterized as double-stranded linear DNA of ∼157 kpb, >94% complete and an average guanine-cytosine (GC) content of 46.4%. The in silico predicted replicative cycle for all three candidates was lytic (Table 1). A multiple alignment of the rearranged genomes showed that the phages shared three syntenic collinear blocks (homologous regions) (Fig. 2), complemented by values >90% of ANI. This allowed us to deduce that our three phages are within the same species. About 190 open reading frames and 200 coding DNA sequences were annotated in each genome. Of the latter, >87% are hypothetical proteins.

FIG. 2.

Progressive multiple alignment performed with MAUVE to compare Klebsiella_virus_K751, Klebsiella_virus_T751, and Klebsiella_virus_T765. Each genome is arranged horizontally with homologous segments (locally collinear blocks) delineated as colored rectangles. Regions inverted with respect to the first genome taken as reference are placed below those that match in forward orientation.

Five common transfer RNAs (tRNAs; tRNA^Thr, tRNA^Met, tRNA^Trp, tRNA^Tyr, and tRNA^Gln), and one additional (tRNA^Asn) in T751 y T765. One small RNA (STnc100) was found in all three genomes. Prokaryotic virulence, toxin-encoding, and antimicrobial resistance genes were not present, demonstrating the absence of risk of transfer by lysogenization or phage-mediated transduction³² (Fig. 3; Supplementary Fig. S5). K751, T765, and T751 clustered with members of the genus Taipeivirus (Fig. 4), as preliminarily expected, as >90% of the raw reads were taxonomically classified within it.

FIG. 3.

Genomic map of Klebsiella_virus_K751. CDS are colored according to their function such as morphogenesis (purple), replication (green), recombination and repair (pink), lysis (light green), tRNA (blue), and other functions (orange). CDS, coding DNA sequences; tRNA, transfer RNA.

FIG. 4.

Phylogenomic tree generated by VICTOR using the complete genome sequences of Klebsiella_phage_T751, Klebsiella_phage_T765, Klebsiella_phage_K751 and members of the nearest Ackermannviridae according to BLASTn. Members of the families Drexlerviridae and Autographiviridae were used as outgroups.

Discussion

The emergence of carbapenem-resistant Klebsiella pneumoniae ST258 strains has led to strong therapeutic challenges for which some alternatives have been proposed. Among these alternatives is phage therapy, however, for many of these potential candidates, the minimum necessary information is not available or is still in the experimental stage. In the context of our research, we isolated three new lytic phages belonging to the family Ackermannviridae that showed significant genetic similarities despite different times and geographical origins.

Our results indicated that our phages belong to the same species. Reports of phages isolated from wastewater against ST258 strains^{7,8,12–14,33} have already shown that there is no exclusivity for a single family or taxonomic group.^30,34,35 Despite the varied phylogenomic distance between families, mostly aquatic environments shared with their host,^36–38 such as wastewater, which due to the presence of human, animal, and hospital waste, are a source rich and varied of these micro-organisms.^39–41

After phenotypic analysis and considering the aforementioned, high stability under different temperature and pH conditions, and a latency period ranging from 10 to 40 min were expected.^14,42–44 Short latency periods and large burst sizes are usually a combination that increases therapeutic potential; however, our values show a discrete difference, which could indicate that phage-host dynamics are being affected possibly due to the inherent facility of K. pneumoniae to generate spontaneous phage-resistant mutants.^{32,34,45–48}

The presence of these mutants increases bacterial density and affects the adsorption rate^49–51 due to poor receptor recognition^47,48,52 resulting in slow adhesion and poor initial diffusion,⁵³ in agreement with our report, in which <30% of free phages were adsorbed in the first 10 min, in contrast to Zurabov and Zhilenkov, who recommend that for the selection of therapeutic phages they should adsorb between 70% and 80% in the same period.^14,44,54 We describe phages of genus Taipeivirus belong to Ackermannviridae family with lytic activity in KPC-producing K. pneumoniae.

After genomic analysis, the mean GC content (46.4%) of our phages was lower than expected for Klebsiella (57.5%), so it could infect other gram-negative bacteria, as previous studies corroborate that genomic GC content accurately predicts (>95%) potential hosts at the phylum level, but not at lower taxonomic levels.^52,55,56 A marked organization of structural and functional genes was evident in each genome, in addition, genes encoding tail proteins (gp5, gp3, and gp17), required for capsid penetration, adsorption,⁵⁷ and irreversible binding to their receptors,^58–60 which have been extensively studied in common models such as enterobacteriophage T4 were also found.⁶¹

Furthermore, five similar and one additional tRNA were present in T751 and T765, which is in agreement with Bailly-Bechet et al. and Maganha de Almeida Kumlien et al.⁶² who state that a lytic phage contains an average of 4 tRNAs and even more, which are useful to compensate the genomic compositional differences with the host, and consequently achieve a more robust integration.^61–63 The presence of tRNAs is possibly unique to lytic phages, as they have not been described in lysogenic phages. The tRNAs present correspond to codons abundant in the phage and probably rare in the host, which gives them an advantage over their competitors by translating their proteins more efficiently, reducing latency time and increasing the burst size.^64–66 In addition, GC content and tRNAs could indicate the closeness between phage and host, although this relationship has not been explored in detail.⁶⁶

Despite reporting encouraging results, phenotypic analysis of our phages revealed certain issues that limit their use as therapeutics and need to be further addressed before proposing them for in vivo or clinical trials.

Conclusion

Three lytic phages of the Ackermannviridae family were recovered against Klebsiella pneumoniae ST258 strains from sewage; however, further characterization is needed for future consideration as therapeutic alternatives. Furthermore, our findings support the strategic targeting of phages for infections by MDR pathogens.

Authors' Contributions

J.A.R. and F.C.M. were responsible for conceptualization. E.T.G. and K.N.R. designed and carried out the methodology, formal analysis, validation, visualization, and research. TEM protocol and observation by K.V. and A.D. The writing—original draft preparation was elaborated by E.T.G. The writing—review and editing article were realized by E.T.G., F.C.M., K.N.R., J.A.R., and L.P. Supervision, project administration, and funding acquisition were conducted by J.A.R., L.P., and I.F. All authors have read and agreed to the published version of the article.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This study is supported by Call for proposals for senior research projects 2019 from Universidad Central del Ecuador, under the Project ID: DI-CONV-2019-052 titled in Spanish: “Alternativa a los antimicrobianos: aislamiento de bacteriófagos a partir de aguas residuales y aplicaciones en el sector salud e industria de alimentos.”

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Figure S5

Supplementary Table S1