Twenty-nine K. pneumoniae strains collected from Renji Hospital, Shanghai Jiao Tong University School of Medicine, and Huashan Hospital, Fudan University (listed in Table 1) were used in this study. All strains were identified by the matrix-assisted laser desorption ionization time-of-flight mass spectrometry (Bruker Diatonic GmbH, Bremen, Germany; D’Andrea et al., 2017). Each K. pneumoniae strain was determined as CRKP with resistance to imipenem or meropenem. The antimicrobial susceptibility was determined using the Kirby-Bauer (K-B) method. The capsular types of these K. pneumoniae strains were determined by wzi gene sequencing according to Brisse et al. (2013). The viable bacterial count was determined using LB agar (1.5% w/v) plates. All strains were cultivated in Luria-Bertani broth (LB, Sangon Biotech, Shanghai, China) at 37^°C and stored at –80^°C in 50% glycerol. K. pneumoniae strain 6570 was used for the phage isolation, the serum assay and the establishment of the infection model.

Phage SH-KP156570 was isolated from a raw sewage collected at Ruijin Hospital, Shanghai Jiao Tong University School of Medicine by using K. pneumoniae strain 6570 as the host bacterium. After centrifugation, the supernatant was filtered with a 0.22 μm filter (Millex-GP Filter Unit; Millipore, United States) and dripped on a double-layer agar plate covered with K. pneumoniae strain 6570. The lytic phage was then purified according to Pires’s method with minor modifications (Pires et al., 2011). The purified phage named SH-KP156570 was stored in SM buffer (100 mM NaCl, 8 mM MgSO₄ 7H₂O, 50 mM Tris–HCl, pH = 7.5) at 4^°C.

The one-step phage growth curve was generated as described previously with some modifications (D’Andrea et al., 2017). Briefly, Phage SH-KP156570 and K. pneumoniae strain 6570 were mixed at an MOI = 0.005. After bathing at 37^°C for 5 min, the mixture was centrifuged, resuspended in 3 mL liquid LB medium. Samples were taken for every 3 min and dripped on the double-layer agar plate at a proper dilution. The number of bacteriophage plaque was counted overnight and the titer of bacteriophage was calculated. The number of plaque-forming units (PFU) was then calculated for 1 mL of the concentrated suspension. The latency period was defined as the time between infection and the shortest incubation time, allowing for the production of phages. The burst size was calculated as the ratio of the final count of released phage particles to the initial count of infected bacterial cells during the latent period. Independent experiments were repeated three times.

Host range analysis of phage SH-KP156570 was performed by the spot test using the strains listed in Table 1. In short, 400 μL log-phase bacterial culture was fully mixed with 4 mL 0.5% agar LB and laid on 1.5% agar LB medium. After the upper agar was solidified, the 5 μL purified phage suspension was spotted onto the plate and incubated overnight at 37^°C to allow plaques to develop.

The genomic DNA of SH-KP156570 was extracted using the standard protocol as previously described (Guo et al., 2017). In brief, the phage in SM buffer was treated with 1 μg/mL DNaseI and RNaseA (Sigma-Aldrich, United States) at 37^°C for 1 h, subsequently 25 mmol/mL ethylenediaminetetraacetic acid. Finally, the concentrated phage DNA was extracted using the UNlQ-10 Column Virus Genomic DNA Isolation Kit (Sangon Biotech, Shanghai, China) and sent to Shanghai Personalbio Biotechnology Co. Ltd for genome sequencing with Illumina high-throughput sequencing platform (Illumina Hiseq 3000). SOAP denovo2 software was used for genome assembly using optimized parameters. Data assembly was proceeding after adapter contamination removing and data filtering by using AdapterRemoval (Lindgreen, 2012) and SOAPec (Luo et al., 2012). The filtered reads were assembled by SPAdes (Bankevich et al., 2012) and A5-miseq (Coil et al., 2015) to constructed scaffolds and contigs. GeneMark was used to predict and analyze the open reading frames (ORFs) of the phage genome. Gene annotation was completed by using the NCBI website¹. Function annotation was completed by blast search against different databases, including NR (Non-Redundant Protein Database; Blake and Cohen, 2001), Gene Ontology (Conesa and Götz, 2008), Kyoto Encyclopedia of Gene and Genomes (Moriya et al., 2007), Cluster of Orthologous Groups of proteins (Powell et al., 2014), and Swissprot. CGview (Stothard and Wishart, 2005) was used to give an overview of the genome information. The results of the HHpred database² were used to locate the gene encoding polysaccharide depolymerase [Uniclust was chosen as the multiple sequence alignment (MSA) generation method, and three was chosen as the maximal number of MSA generation steps].

The ORF41 gene was amplified from the DNA of the purified phage SH-KP156570 using primers 6570-ORF41-F (5′-CAGCAGCAGACGGGAGGATCCATGTCCACGATTACACA ATTC-3′) and 6570-ORF41-R (5′-CTCGAGTGCGGCCG CAAGCTTTTAGTTACTTCTCTCTTCAGC-3′). The 2292-base PCR amplification product was cloned into the N-terminal 6 × His labeled pSUMO3 expression vector (LifeSensors, United States) via BamHI and HindIII sites (New England Biolabs). The recombinant plasmid was verified by DNA sequencing and transformed into E. coli BL21 (DE3). The BL21 cells were cultivated to OD₆₀₀ = 0.8. Then, the BL21 cells were induced with 0.1 mM isopropyl-β-D-thiogalactopyranoside (IPTG, Sangon Biotech, China). The mixture was centrifuged and the supernatant was discarded. The sediment was resuspended with lysis buffer [20 mM Tris–HCl (pH = 8.0), 0.5M NaCl, 10% glycerol] and phenylmethylsulfonyl fluoride (PMSF, Sangon Biotech, China) to a final concentration of 0.1 mg/mL. After centrifuged, the supernatant was collected. The Ni-NTA column (GE Healthcare, United States) was prerinsed using lysis buffer. After protein binding to Ni-NTA column, it was eluted with a gradient of 20 and 40 mM imidazole, and finally eluted with 300 mM imidazole. Then digested by SUMO protease (LifeSensors, United States). The purified depolymerase was confirmed by SDS-PAGE electrophoresis and named K19-Dpo41.

Klebsiella pneumoniae strain 6570 was cultured in fresh TSB medium at 37^°C for 5 days. The purification of CPS was performed according to the method from previous studies with some modifications (Gorodnichev et al., 2021; Li J. et al., 2021). Firstly, 60 μL of formaldehyde solution (36.5%) was added to 10 mL of bacterial culture and incubated 100 rpm at room temperature for 1 h. Subsequently, 1M NaOH was added to the system, with agitation at room temperature for 3 h. Cell suspensions were centrifuged at 16,800 g for 1 h at 4°C (Beckman, JA-25.50, United States). The supernatant was filtered through a 0.22 μm filter and dialyzed overnight in ddH₂O with a 12–14 KD MWCO membrane (Thermo Scientific, United States). Then, the cationic detergent cetyltrimethylammonium bromide was added (0.5% w/v) to precipitate polysaccharide. Dissociate with different concentrations of CaCl₂ and centrifuge the supernatant. Selective precipitation with 20 and 50% ethanol and washed with NaCl. Next, the sample was subjected to Capto adhere chromatography after ultrafiltration with a 30 KD membrane. The final sample was filtered through a 0.22 μm filter. Finally, the CPS obtained by dialysis was dried and weighed.

The activity of K19-Dpo41 was assessed by spot test (Khan Mirzaei and Nilsson, 2015). After K. pneumoniae strain 6570 was poured onto a LB agar, a serial 5 μL polysaccharide depolymerase in different concentration was spotted on the plate and incubated overnight at 37°C. The K19-Dpo41 activity was monitored by observing the formation of translucent halo zones. In addition, the sensitivity of other K. pneumoniae strains to K19-Dpo41 was also determined by single-spot assay (Wu et al., 2019).

Size exclusion chromatography-High performance liquid chromatography (SEC-HPLC) was used to determine the depolymerase activity of K19-Dpo41 on the CPS. The purified CPS was dissolved in 50 mM Na₂HPO₄ (pH = 7.0) to a final concentration of 0.5 mg/mL and incubated with K19-Dpo41 (10 μg/mL) or ddH₂O at 37^°C for 30 min. The reactions were terminated by heating at 100°C for 10 min. The mixture was analyzed by a HPLC system (Waters, Milford, MA, United States) equipped with a TSKgel G5000 PWxl column (inner diameter 7.8 mm × 300 mm; Tosoh Corporation Bioscience, Tokyo, Japan) and a refractive index detector (Waters, Milford, MA, United States). The column was run with the mobile phase of the phosphate-buffered saline (PBS, pH = 7.0) at 1 mL/min.

The serum resistant assay was performed as previous studies (Liu et al., 2020; Li J. et al., 2021). K. pneumoniae strain 6570 at a concentration of 5 × 10⁴ CFUs were suspended in PBS, treated with either K19-Dpo41 or K64-ORF41 (a depolymerases specific to K. pneumoniae K64 serotype), respectively, to a final concentration of 100 μg/mL. After a pretreatment at 37°C for 1 h, the suspension was mixed at a 1:3 v/v ratio with active or heat-inactivated human serum obtained from healthy volunteers. A control without depolymerase was also set up for significance. The mixture in final volume of 100 μL was incubated for 1 h at 37^°C, and then 5 μL aliquots were removed, diluted and cultured on 1.5% LB agar plate for colony enumeration. Each test was performed at least in three independent experiments. Comparisons between any two experimental groups were made by the two-sample Student’s t-test. P < 0.05 was considered statistically significant.

Wax moth larvae (G. mellonella) were obtained from Tianjin Huiyude BioTech, China. The operating procedure was carried out in accordance with the previous studies (Majkowska-Skrobek et al., 2016; Oliveira et al., 2019; Gorodnichev et al., 2021; Shahed-Al-Mahmud et al., 2021). In short, K. pneumoniae strain 6570 from overnight culture were grown in fresh LB at 37°C to exponential phase, harvested (10,000 rpm, 10 min), washed with PBS and suspended in PBS to a concentration of 5 × 10⁶ CFU/mL. Larvae were inoculated with 10 μL bacterial suspension containing 5 × 10⁴ CFUs into the last proleg using 25 μL syringe. The anti-virulence effect of enzyme on K. pneumoniae strain 6570 was estimated by inoculated larvae with either K19-Dpo41 administered at 5 min or 30 min after bacterial infection. Larvae injected with either PBS or 5 × 10⁴ CFUs of K. pneumoniae strain 6570 were included as control groups. After inoculation, caterpillars were kept at 37°C in the dark for 5 days. The results were expressed as the percentage survival rate estimated on the basis of the touch-provoked motility at each day post injection. For each option, at least three independent experiments were performed (10 larvae per trial). The data was presented as the average of three experiments.

GraphPad Prism (version 8, GraphPad Software, United States) software was used for the statistical analysis. The data were tabulated as mean ± SD. In the serum resistance analysis, comparisons between any two experimental groups were made by the two-sample Student’s t-test. In the Galleria mellonella larvae infection model, survival curves were plotted using the Kaplan–Meier method, and the survival analysis was performed by using the log-rank Mantel-Cox test. P < 0.05 was considered to be statistically significant.

The strain specimens were collected with the written and informed consent of the patients. The conduct and procedures involved in the present work were approved by the Ethics Committee of Shanghai Jiao Tong University School of Medicine.

All 29 clinical isolates were identified as CRKP and resistant to a variety of antibiotics, including cefuroxime (CXM, 28/29), amikacin (AMK, 15/29), piperacillin-tazobactam (TZP, 28/29), levofloxacin (LVX, 27/29), ciprofloxacin (CIP, 27/29), ceftazidime (CAZ, 29/29), cefoperazone sodium Shubatan (CSL, 27/29), cefepime (FEP, 29/29), Imipenem (IPM, 19/29), meropenem (MEM, 28/29), trimethoprim-sulfamethoxazole (SXT, 17/29), and aztreonam (ATM, 27/29). Twenty-nine K. pneumoniae were genotyped by wzi sequencing. Twenty of them were Type K19, others were K1, K2, K20, K47, K57, and K64 (Table 1).

Klebsiella pneumoniae strain 6570 and the sewage collected from hospital were co-cultured overnight and phage SH-KP156570 was successfully isolated. The phage was spread over K. pneumoniae strain 6570 to form a double-layer agar plate and cleared plaques formations and faint halos were observed. The size of the faint halos increased over time (Figure 1A), which we suspected indicated the presence of a phage-derived depolymerase based on our previous study (Li J. et al., 2021). To understand the relationship between SH-KP156570 and capsular types, we performed a host spectrum analysis of SH-KP156570 on all 29 strains using spot tests. Faint halos and plague were observed over time for all twenty K19-type strains, while absences of plaque or halo were observed for nine non-K19-type strains (Table 1). The results indicated that phage SH-KP156570 targeted K19 CPS exclusively.

The life cycle of SH-KP156570 was revealed with a one-step growth curve. The phage had a short latent period of 6 min followed by a rise period of the phage progeny until a stationary phage was reached at 9 min (Figure 1B). The burst size was about 80 PFU per infected cell.

A single contiguous sequence of 38,667 bp was de novo assembled from 8,102,358 raw sequencing reads with more than 30,330 × coverage. Hence, the phage SH-KP156570 had a genome size of 38,667 bp with a GC content of 50.85%. The whole genome sequence of the SH-KP156570 was deposited in National Genomics Data Center³ with the accession number (GWHBGZN00000000). We identified 47 ORFs. Genes encoding putative functional protein were assigned to 29 of the 47 ORFs (61.7%). The average size of these ORFs was 851 bp. Phage-encoded proteins could be divided based on their functions: DNA packaging and morphology-related proteins (6), DNA replication/recombination/modification proteins (16), host lytic proteins (4), and unknown proteins (3). The remaining genes were encoded as hypothetical proteins. Among the 29 functional annotated genes, genes involved in DNA processing and packaging machinery, structural and lytic proteins, DNA replication, DNA recombination, and repair molecular mechanisms were identified. No putative integrase genes, virulence factors, toxins, or antimicrobial resistance genes were found in the genome of SH-KP156570 (Figure 2A).

Gene ORF41 (2292bp), coding for the tail fiber protein, was predicted to be a depolymerase-encoding gene. The N-terminal conserved domain (11 to 113 aa) of ORF41 protein showed a high similarity with the N-terminal sequence of T7 phage, while the central region (252 to 296 aa) showed a high similarity with the pectate lyase (Figure 2B). This ORF protein contained a β-helical pectin lyase domain, and had a 28% identity to the tail spike protein TSP3 from Escherichia coli phage CBA120 (a lipopolysaccharide lyase of E. coli; Greenfield et al., 2019). These results suggested that the ORF41 protein might be a polysaccharide depolymerase in phage SH-KP156570.

The ORF41 gene of phage SH-KP156570 was cloned into the pSUMO3 expression vector. The recombined K19-ORF41-pSUMO3 was expressed and purified by Ni-NTA column. The purity of the purified K19-Dpo41 (about 84 kDa) was more than 95% confirmed by SDS-PAGE gel analysis (Figure 3A). The purified recombinant depolymerase K19-Dpo41 were diluted to different concentrations (4.2 μg/mL∼0.42 mg/mL) for the spot test to assess the polysaccharide depolymerization activity. Clear halos were observed at different K19-Dpo41 concentrations as low as 4.2 μg/mL (Figure 3B). The SEC-HPLC results confirmed the depolymerization of K19-Dpo41 against the CPS of K. pneumoniae strain 6570. The untreated CPS showed a single peak with a retention time from 12 to 16 min, however, after incubation with K19-Dpo41 at 37°C for 30 min, the peak of CPS disappeared, demonstrating that CPS was degraded by K19-Dpo41 (Figure 3C).

The activity of K19-Dpo41 was further tested on all 29 K. pneumoniae strains using the spot assay. On plates incubated with K19-type K. pneumoniae strains, the recombinant protein K19-Dpo41 generated a translucent halo zone (Table 1). Meanwhile, no translucent spots formed on any non-K19-type K. pneumoniae strains as expected. The results showed that depolymerase K19-Dpo41, same as the parental phage SH-KP156570, exhibited specificity to K19-type K. pneumoniae.

We assessed the bactericidal effect of serum after depolymerase treatment on K. pneumoniae strain 6570. After 1 h of incubation, the survival rate of bacteria treated with K19-Dpo41 decreased by around 70% compared with those of untreated bacteria and bacteria treated with K64-ORF41, a depolymerase specific for K64-type K. pneumoniae, demonstrating that K19-Dpo41 could improve the susceptivity of K19 K. pneumoniae to serum killing (Figure 4A).

We also evaluated the anti-k. pneumoniae infection efficacy of K19-Dpo41 using K. pneumoniae infected G. mellonella larvae model. Our results showed that in the K. pneumoniae infection group, 90% of the larvae injected with K. pneumoniae strain 6570 died within 3 day. In contrast, a single dose of K19-Dpo41 given 30 min after bacteria injection increased the survival rate of the larvae to 50%, and the same dose given 5 min reached to 70%. No mortality of larvae was observed in the group injected with PBS.