The clinical strains P. aeruginosa, Acinetobacter baumannii, E. coli, K. pneumoniae, and Staphylococcus aureus, isolated from Ruijin Hospital (Shanghai, China), are shown in Supplementary Table S1. All of these isolates were recorded in a computerized database that included source and antimicrobial data. All the bacteria were routinely cultured in LB medium. E. coli DH5α and E. coli BL21(DE3) were grown in LB broth at 37°C. When needed, 50 ng/ml of kanamycin was added to the growth medium.

Phage particles were isolated from sewage obtained from Ruijin Hospital (Shanghai, China) by five rounds of plaque purification. Phage JD010 were purified by precipitation with PEG 8000, followed by cesium chloride (CsCl) density-gradient ultracentrifugation. The morphology of phage JD010 was examined after staining with 2% phosphotungstic acid and using a Hitachi 700 transmission electron microscope. The phage DNA was extracted as described elsewhere (Lin et al., 2010) and sequenced with an Illumina MiSeq platform. Using the BLAST at the NCBI¹, comparative genome analysis of phage JD010 was carried out; the prediction of the conserved protein domain was conducted using BLASTP and the NCBI Conserved Domain Database².

lysPA26 was amplified with the primers lysPA26-F/lysPA26-R (the BamHI/XhoI restriction enzyme sites are shown in bold in the primer sequences in Supplementary Table S2) with JD010 phage genomic DNA as the template. To construct the LysPA26 expression plasmid, the PCR product and vector pET28b were digested with BamHI and XhoI restriction enzymes (Fermentas, USA), then the PCR restriction fragment was ligated into the pET28b vector (Supplementary Table S2) at the corresponding restriction sites. After ligation, pET28b-lysPA26 plasmid was used to transform E. coli BL21(DE3), and expression was induced with 1 mM isopropyl-β-D-thiogalactopyranoside for 5 h at 25°C.

Antimicrobial activity was determined as described previously (Rodriguez et al., 2011) with some modifications. Briefly, P. aeruginosa D204 was grown in LB broth at 37°C to exponential phase. Bacterial culture was harvested at 4000 g for 5 min and washed once with dilution buffer (20 mM Tris-HCl, pH 8.0). Then, the cell pellet was resuspended in dilution buffer and adjusted OD₆₀₀ to 1.0. For the antimicrobial activity assay, different concentrations of LysPA26 were added into P. aeruginosa D204 bacterial suspension with or without OMPs (EDTA, Triton X-100, or CHCl₃). After 30 min of incubation at 37°C, the mixture was serially diluted and plated. To determine the survival rate, residual viable cell numbers (CFUs) on the plate were measured after incubation at 37°C for 24 h. For the negative control, the same volume of buffer was added instead of endolysin LysPA26. The antibacterial activity was expressed as the decrease in viable bacterial counts. Alternatively, the antibacterial activity was presented as the relative inactivation in logarithmic units [= log₁₀(N₀/N_i), where N₀ = the number of residual cells (in the negative control) and N_i = the number of viable cells counted after incubation with endolysin]. All experiments were performed in triplicate.

To study the effect of temperature on the activity of LysPA26, 50 μg LysPA26 was added into 100 μl test bacteria suspension, and the mixture was incubated for 30 min at different temperatures. For the thermal stability assay of LysPA26, the endolysin was incubated at 100°C for 10 min, then the heat-treated LysPA26 was added into the P. aeruginosa D204 bacteria suspension with or without 1 mM EDTA. The antibacterial activities were measured as described above. To evaluate the effect of pH on lytic activity, 10 μl LysPA26 (5 mg/ml) was incubated with 90 μl buffers with different pH ranges (50 mM sodium acetate for pH 4.0–6.0 and 20 mM Tris-HCl for pH 7.0–11.0) for 30 min. The effect of saline concentrations on the lytic activity of LysPA26 was tested by adding different NaCl concentrations to cell suspensions for 30 min in 20 mM Tris-HCl buffer, pH 7.0, at 37°C. The residual activity of each treatment group relative to the activity of the control group (100% activity) was determined. All experiments were performed in triplicate.

To test the lytic zymogram range of LysPA26, P. aeruginosa, A. baumannii, E. coli, K. pneumoniae, and S. aureus isolates were challenged. All isolates, in mid-exponential phase, were washed and resuspended in dilution buffer (20 mM Tris-HCl, pH 8.0), 50 μg LysPA26 was incubated with the various bacterial species isolates in a final volume of 100 μl, and then serial dilutions of the samples were plated for CFU counting after incubation for 30 min; antibacterial activities were measured as described above. All experiments were performed in triplicate.

The biofilm eradication assay was carried out by crystal violet staining as described previously (Sass and Bierbaum, 2007). MDR P. aeruginosa strain 8328 was cultured in the wells of a polystyrene 96-well plate (BD Falcon) at 37°C for 48 h to allow biofilm formation. The planktonic cells in the culture were discarded and the plate washed with phosphate-buffered saline three times. LysPA26 (100 μg in 200 μl) was added to the plate wells, incubated at 37°C for 2 h, and then washed twice with phosphate-buffered saline. Crystal violet (100 μl 1%) was added and incubated for 30 min; subsequently, 33% acetic acid was added to dissolve the stain, and the optical density was measured at OD₆₀₀. The elimination activity of LysPA26 was also measured by enumerating the reduction of the residual live biofilm cells (Fu et al., 2010). Briefly, after discarding planktonic cells, the biofilm cells were treated with LysPA26 for 2 h. The contents of the treated biofilm-containing wells were mixed fully with a pipetting device, followed by sonication at 42 kHz in a water bath sonicator to make the biofilm cells become planktonic cells. For each experiment, samples were analyzed in triplicate.

Data were analyzed by a Student’s t-test, and a value of P < 0.05 was considered statistically significant.

A phage, named JD010, was isolated from sewage water in Ruijin Hospital (Shanghai, China). Based on electron microscopy, phage JD010 was classified as a member of family Podoviridae with 15 nm tail length and 50 nm head width (Figure 1A). Phage JD010 DNA was obtained and sequenced. Phage JD010 has a double-stranded DNA genome of 50609 bp containing 75 ORFs, with a G+C content of 55.5% (Figure 1B). BLAST analysis showed that the JD010 genome had sequence similarity with Pseudomonas phage PA11, with 89% identity and 83% coverage. ORF 26 of the phage JD010 genome was predicted as the putative endolysin and named LysPA26. BLAST analysis revealed LysPA26 was homologous to the endolysin (ORF033) of Pseudomonas phage PA11, with 99% amino acid identity.

LysPA26 was predicted to belong to the lysozyme-like domain family of superfamily cd00442, which work as peptidoglycan hydrolases. A comprehensive bioinformatics study showed that LysPA26 was composed of a single conserved sequence motif (lysozyme domain), involving the sequence from the 6th to 138th amino acids of the 145 amino acids. The catalytic residues (active sites) of LysPA26 were predicted to be E13 (Glu) and T28 (Thr) (Figure 2A). Recombinant LysPA26 was overexpressed and successfully purified from the soluble fraction (Figure 2B).

To verify the antibacterial activity of LysPA26, P. aeruginosa D204 was challenged with different concentrations of LysPA26. LysPA26 showed efficient bactericidal activity against exponentially growing P. aeruginosa D204, and its antibacterial activity was enhanced as the concentration increased (Figure 3). It was interesting that LysPA26 could kill P. aeruginosa D204 without an OMP. The lytic ability of LysPA26 did not increase any further when the concentration of LysPA26 was more than 0.5 mg/ml. Thus, the endolysin concentration of 0.5 mg/ml was used in the following assays.

With pre-treatment by OMPs, such as EDTA, Triton X-100, or CHCl₃, lysin can effectively destroy Gram-negative bacteria as reported. In order to test the effect of OMPs on the bactericidal activity of LysPA26, EDTA, Triton X-100, and CHCl₃ were selected to individually pretreat P. aeruginosa D204, and antibacterial activity was expressed as the decrease in viable bacterial counts (Figures 4A–C). Different concentrations of OMPs were tested to determine whether they had synergistic effects on the antimicrobial activity of LysPA26. The results showed the EDTA could enhance the lytic ability of LysPA26 against P. aeruginosa D204 in a certain range, but Triton X-100 and CHCl₃ could not. Thus, 1 mM EDTA was selected for use as the OMP in the following experiments, where needed.

The optimal conditions for the bactericidal activity of LysPA26, such as temperature, pH, and saline concentration, were tested. Temperature was the key factor affecting the bactericidal activity of endolysin. As shown in Figure 5A, LysPA26 exhibited high antibacterial activity from 37 to 50°C. The antibacterial activity of LysPA26 decreased when the temperature was lower than 25°C or higher than 60°C. As for the effect of saline concentration on the activity of LysPA26, NaCl was selected for the test. The results showed that a NaCl concentration below 150 mM did not show a significant effect on the activity of LysPA26. However, when the concentration was high enough, up to 300 mM, it could reduce the efficacy of LysPA26 (Figure 5B). As for the effect of pH on activity of LysPA26, it was efficiently active from pH 6.0 to 8.0, and the optimal pH was 8.0 (Figure 5C). The results showed the LysPA26 has a good pH stability in the antibacterial activity assay.

We also checked the thermostability of LysPA26, it was heat-treated at 100°C for 10 min before being used to challenge exponentially growing P. aeruginosa D204; the cells were treated with EDTA or were untreated. For the challenged bacteria without EDTA treatment, the activity of LysPA26 decreased to less than 20%. However, when treated with the EDTA, LysPA26 still maintained over 90% of its activity (Figure 5D). This may be attributed to the stability of LysPA26 catalytic sites and the reliable permeabilization of EDTA, which makes the peptidoglycan fully exposed.

To further estimate the zymogram range of LysPA26 against clinical MDR strains, clinical samples including Gram-negative (A. baumannii, K. pneumoniae, P. aeruginosa, and E. coli) and Gram-positive (S. aureus) bacteria were tested. LysPA26 exhibited high antibacterial activity against the tested Gram-negative bacteria (Figure 6). It was apparent that the antibacterial spectrum of LysPA26 was broader than phage JD010. Of note, LysPA26 could not kill P. aeruginosa A2210, which is the host of JD010. A methicillin-resistant S. aureus isolate was tested to detect the antibacterial activity of LysPA26 against Gram-positive bacteria. The results showed no antibacterial activity against the strain.

Pseudomonas aeruginosa is one of the most common biofilm-forming bacteria. We designed experiments to test whether LysPA26 had the ability to destroy P. aeruginosa biofilm. P. aeruginosa 8328 was attached to a plate for 48 h to form biofilm before adding LysPA26. We found that the LysPA26 had the ability to efficiently destroy existing biofilm cells. The elimination ability of biofilm was dose dependent; the optical density (OD₆₀₀) staining of the biofilms was reduced significantly upon addition of LysPA26 up to 50 μg (Figure 7A). Approximately 1–2 log in the number of viable counts of biofilm cells in the plate wells was disrupted by LysPA26 (100 μg) (Figure 7B). Combined with the other results, the elimination efficiency for the biofilm cells was comparatively lower than that for the planktonic cells.