Our previous study identified a K1 capsule depolymerase (K1-ORF34) from NTUH-K2044-K1-1 (a K1-specific phage; accession number AB716666) (27). In order to isolate a K2-capsule-degrading enzyme, phage capable of infecting the K. pneumoniae 1611E strain (capsular type K2) was isolated from untreated water as clear plaque with translucent halo (designated 1611E-K2-1). The sensitivity of capsular type K2 phage 1611E-K2-1 was evaluated against 7 other capsular-type-K2 strains, whose capsular types were determined by PCR using wzy primers. Results showed that phage 1611E-K2-1 could infect all capsular type K2 strains.

The full genome of phage 1611E-K2-1 was determined to be 47,797 bp in length (accession number MG197810). Annotation of the genome sequences predicted that this phage should contain 17 open reading frames (ORFs) of more than 500 bp. Our analysis on the ORFs of phage 1611E-K2-1 revealed that the predicted ORF16 exhibited 46% amino acid identity with a tailspike 63D sialidase, suggesting that the protein might correspond to a capsule depolymerase. Subsequently, the K2-ORF16 gene was cloned and expressed in E. coli, and the purity of the purified recombinant K2-ORF16 protein was verified ( Figure 2A ). Upon spotting a plate inoculated with top agar containing K. pneumoniae 1611E, the recombinant K2-ORF16 protein generated a translucent spot resembling the plaque halo ( Figure 2B ). To confirm the putative depolymerase activity of the purified protein, CPS was extracted from 1611E and then treated in vitro with K2-ORF16 protein; the resulting products were separated via SDS-PAGE and stained with Alcian blue ( Figure 2C ). The results showed that the K2-ORF16 protein degrades the K2 CPS. The specificity of this enzyme for capsular type K2 was further confirmed in 7 other capsular-type-K2 strains. These results demonstrated that the K2-ORF16 protein is a K2 capsule depolymerase.

To verify the CPS modifications remain intact, K1 CPS treated by K1-ORF34 capsule depolymerase was digested and further analyzed. After incubation with purified K1-ORF34 protein, K1 CPS was reduced to oligosaccharides. Mass spectrometry revealed that the majority of oligosaccharides were hexa-saccharide (m/z 1145), contained dimer of a structural trisaccharide repeating unit: (→3)-β-D-Glc-(1→4)-[2,3-(S)-pyruvate]-β-D-GlcA-(1→4)-α-L-Fuc-(1→), and had the unusual feature of extensive pyruvation of glucuronic acid and acetylation of fucose at C2-OH or C3-OH with nonasaccharide (trimer) and trisaccharide (monomer) as minor components ( Supplementary Figures 1A-C ). MS/MS analysis revealed the molecular weight of the oligosaccharide is 1145, suggesting that glucuronic acid was reduced to a double-bond derivative that absorbed wavelength of 232 nm. The major enzymatic product is a hexasaccharide 1: D-GlcA*-(1→4)-α-L-Fuc-(1→3)-β-D-Glc-(1→4)-[2,3-(S)- pyruvate]-β-D-GlcA-(1→4)-α-L-Fuc-(1→3)-β-D-Glc; GlcA*: 4,5-unsaturated glucuronic acid (delta 4,5 GlcA). NMR spectra (δ 1.6, s and δ 1.55, s for CH₃; δ 2.1, s and δ 2.05, s for OCOCH₃) indicated that the pyruvate and acetyl modifications were still present ( Supplementary 1D ). These results suggested that the K1-ORF34 enzyme belongs to the lyase class of enzymes and can be used in the preparation of K1 carbohydrate antigens for the development of K1 CPS-conjugated vaccine.

To verify the CPS modifications remain intact, K2 CPS treated by K2-ORF16 capsule depolymerase was digested and further analyzed. The spectra showed that the major components of K2-ORF16-digested CPS were tetrasaccharide monomer 2 (m/z 703, Man-[(3←1)-α-GlcA]-(1→4)-α-Glc-(1→3)-β-Glc) and the minor components were dimer (octasaccharide) and trimer (dodecasaccharide) ( Supplementary Figures 2A–C ). The product with m/z of 703 was further examined by ESI-MS/MS analysis. The spectrum showed that this product contained three hexoses and one uronic acid, consistent with the previously reported structure of K2 CPS (28). We had assigned the NMR spectra and compared the reported K2 CPS spectra (28). Based on the information, the ¹³C chemical shift at C-4 of mannose was assigned as δC 73.3. In our NMR studies, the up-field 13C chemical shift at the same position of the oligosaccharide was found to be δC 66.1, suggesting the non-reducing end was located at C-4 position of mannose ( Supplementary Figures 2D–F and Supplementary Table 1 ). These results indicated that the K2-ORF16 enzyme belongs to a kind of hydrolase and can apply for the preparation of K2 carbohydrate antigens for the development of K2 CPS-conjugated vaccine.

The MS analysis showed that the digested K1 and K2 oligosaccharides used for conjugation consisted primarily of hexasaccharides 1 (two repeat units of K1 CPS) and tetrasaccharides/octasaccharides 2 (one/two repeat unit of K2 CPS), respectively. The compounds 1 and 2 were subjected to the Kochetkov amination using ammonium carbonate to change the anomeric hydroxyl group to the amine group; then, the glycosyl amines were reacted with DTSSP linker to form the glycosyl disulfides. After DTT treatment, the K1/K2 digested CPS-SH was reacted with CRM197-maleimid to form the K1 and K2 CPS-conjugated vaccine candidates ( Scheme 1 ). The epitope ratio (K1 = 5.38, K2 = 7.42, sugar/protein) was determined by MALDI-TOF mass spectrometry ( Supplementary Figures 3 , 4 ).

The body weight ( Figure 3A ) and rectal temperature measurements ( Figure 3B ) of mice were recorded right before and one day after each vaccination of K1 CPS-conjugated vaccine. Sera of mice were collected one week after the 1^st, 2^nd, and 3^rd administrations of K1 CPS-conjugated vaccine, and blood levels of liver (ALT) and renal (BUN and creatinine) markers were determined ( Figure 3C ). The mouse body weight measurements from before vaccination and one day after vaccination did not have significant difference; rectal temperature measurements for mice administered with K1 CPS-conjugated vaccine were all in the normal range. The blood level measurements of liver and renal markers were also all in the normal range and did not alter in vaccinated mice compared to control mice. These results suggested no side effects in mice administered with three doses of K1 CPS-conjugated vaccine. Observations were similar in K2 vaccinated mice, except no serum biochemical tests were done.

Sera of mice that were administered with three doses of K1 or K2 CPS-conjugated vaccine by IM injection were collected for assessments of antibody production and bactericidal activity. The results showed antibodies induced by the K1 or K2 CPS-conjugated vaccine interacted with their original CPS ( Figure 4A, C ). The major induced antibody of K1 CPS-conjugated vaccine was IgG1 ( Supplementary Figure 5 ). The persistence of antibodies was still observed 9 months after administration of the last vaccine booster dose ( Figure 4B, D ).

The bactericidal activity assay results revealed that the K1 bacteria bactericidal titer of sera from mice immunized with K1 CPS-conjugated vaccine was 32~128, which means sera from mice immunized by K1 CPS-conjugated vaccine killed more than 50% of K1 bacteria at a dilution of up to 1: 128 (i.e., titer of 128). On the other hand, sera from non-immunized mice did not exhibit appreciable bactericidality even at the lowest (1: 2) tested dilution. For the K2 bacteria, the bactericidal titer of sera from mice immunized with K2 CPS-conjugated vaccine was 32~64 ( Figure 5 ).

Notably, cross-capsular-type protection was not observed. Comparisons of immunization with K1 or K1 CPS conjugated vaccine. The serum bactericidal activities were higher in conjugated vaccines ( Figures 6A, B ). Therefore, we can conclude that the CPS-conjugated vaccines successfully induced production of capsular-type-specific antibody in mice with promising bactericidal activity.

The immunized mice were challenged with K. pneumoniae to evaluate its in vivo protective effect. Experimental mice were inoculated with K1 or K2 CPS-conjugated vaccine by IM injection once a week for 3 weeks; the control animals were administered with adjuvant only on the same schedule. One week after the third vaccination, the mice were infected with 1×10⁴ CFU K. pneumoniae NTUH-K2044 ( Figure 7A ) or NTUH-A4528 per mouse by IP injection ( Figure 7B ). Observation over the 30 days after the K. pneumoniae challenge revealed significant higher survival rate of the mice that received the K1 or K2 CPS-conjugated vaccine than that of the mice that received adjuvant only.

Mice received K1 CPS-conjugated vaccine were challenged with K. pneumoniae NTUH-K2044 at nine months after the third vaccination ( Figure 7C ). The results indicated that the vaccine still has effective protection nine months after the inoculation. Both groups were protected from pyogenic liver abscess and mortality.

K1 and K2 are the prevalent capsular types of K. pneumoniae strains and cause invasive infections. Therefore, the protection efficacy of a divalent vaccine (a mixture of equal amounts of K1 and K2 CPS-conjugated vaccines) was further examined ( Figures 8A, B ). Mice infected with 1 × 10⁴ CFU of NTUH-K2044 or NTUH-A4528 by IP injection resulted in 80% and 100% death among adjuvant-immunized (control) mice, respectively. In contrast, no mortality was observed among divalent-vaccinated mice after IP injection. Thus, we can conclude that the divalent vaccine effectively protected mice from K1 or K2 K. pneumoniae infection.

The agar overlay method was used for isolation of pure phage preparations and titer determinations of the phage. The infectivity of the phage was determined by spot test (27). Phage genomic DNA was extracted using Qiagen Lambda kit following manufacturer’s protocol with modifications as follows (27). After the phage was precipitated and lysed, phage DNA was extracted by phenol/chloroform and precipitated with ethanol. The genome sequences of the phage DNA were determined by Illumina GAII sequencing. The complete genome sequences were deposited to GenBank. The full genome of phage NTUH-K2044-K1-1 and phage 1611E-K2-1 were deposited in GenBank under accession number AB716666 (https://www.ncbi.nlm.nih.gov/search/all/?term=AB716666) and MG197810 (https://www.ncbi.nlm.nih.gov/search/all/?term=MG197810), respectively.

The recombinant His-tagged K1-ORF34 protein was expressed and purified as described in our recent study (27). The K2-ORF16 gene was amplified by PCR using primers 1611E-ORF16-F (5’-CAAACATCACGGTGACGCTAGCATGACCATTATCAAACG-3’) and 1611E-ORF16-R (5’-CTTTTAACATTTAGCACTCGAGTGTAAAATTAATAATG-3’); then, the product was digested with NheI and XhoI. The K2-ORF16 fragment was cloned into NheI and XhoI double-digested pET28c plasmid (Novagen). The resulting K2-ORF16-pET28c plasmid was transformed into E. coli BL21 (DE3). The recombinant His-tagged K2-ORF16 protein was expressed by inducing with 0.4 mM IPTG overnight at 16°C, and the recombinant protein was purified by nickel affinity following manufacturer’s protocol (Qiagen).

In order to eliminate the contamination of CPS with O polysaccharide, the K1 and K2 CPSs were purified from K. pneumoniae NTUH-K2044 ΔwbbO and NTUH-A4528 ΔwbbO mutant strains (respectively) and quantified by phenol-sulfuric acid method as previously described (47).

The enzymatic activity of K1 lyase towards the substrate K1 CPS was evaluated by monitoring elevated absorption at 232 nm, performed at 25°C in an Ultrospec 4000 UV/Visible spectrophotometer (Amersham Pharmacia Biotech, Piscataway, NJ, USA) using a 1 mL cuvette. After K2 enzymatic reaction, the reaction supernatant was reacted with equal volume of DNSA (20 mg/mL in 0.7 M NaOH), boiled for additional 5 min, cooled on ice, and determined by monitoring the absorption at 535 nm.

The reaction for oligosaccharide production used 50 mg CPS and 100 μg enzyme in 20 mM Tris-HCl and 300 mM NaCl in pH 8 at 37°C for 4 hr. After denaturation of the enzyme, the mixture was centrifuged, and the supernatant part was separated by Biogel P6 column (1 cm D x 13 cm H) and eluted in ddH₂O condition. The pure oligosaccharides were lyophilized for glycoconjugation.

The chemical structures of K. pneumoniae NTUH-K2044 (K1) and NTUH-A4528 (K2) CPS have been defined previously (26, 28). The lengths and modifications of enzyme-digested K1 and K2 CPS were determined by mass spectrometry and NMR as described in the Supplementary Methods .

The digested K1 and K2 CPSs were separately conjugated with CRM197 carrier protein as described in the Supplementary Methods . In brief, glycosyl amines were obtained by reducing CPS via Kochetkov amination. A thiol linker was attached to the resulting K1/K2 digested CPS-SH. Then, the linker-modified products were reacted with CRM197-maleimid (Scheme 1). The resulting digested K1/K2 CPS-CRM197 conjugates were characterized by MALDI-TOF analysis to confirm the incorporation of carbohydrate.

All animal experiments followed the guidelines in the Handbook of Laboratory Animal Care of the National Laboratory Animal Breeding and Research Center, National Science Council of Taiwan and were approved by the Institutional Animal Care and Use Committee of the National Taiwan University College of Medicine.

The CPS-conjugated vaccine was diluted to 100 μg/mL in phosphate buffered saline (PBS), and the glycolipid adjuvant C34 was dissolved to 100 μg/mL in Dimethyl sulfoxide (DMSO) (48). Mice were injected intra-muscularly (IM) with 100 μL of a mixture of CPS-conjugated vaccine (2 µg K1, 2 µg K2, or 2 µg K1 plus 2 µg K2) and 2 µg adjuvant, and the booster shots were given by IM injections after one and two weeks of the inoculation. Then, sera were collected from punctured submandibular vein.

Purified CPSs were subjected to serial two-fold dilutions, and different amounts (ranging from 2000 ng to 7.8125 ng per spot) were transferred onto membrane by dot blotting. After incubating the dot blot with sera (1:1000 dilution) from vaccinated mice, the blots were then developed with chemiluminescent substrates per the manufacturer’s protocol (Millipore) and exposure to X-ray films for antibody detection.

Aliquots of mouse sera were inactivated by heating at 56 °C for 30 min and then subjected to two-fold serial dilution (1/2 to 1/256) with normal saline. Aliquots of 25 μL of diluted serum and 12.5 μL of fresh-cultured NTUH-K2044 or NTUH-A4528 bacterial suspension (100 CFU) were combined and incubated at 37°C for 15 minutes. After incubation, 12.5 μL of newborn rabbit complement (Pel-Freez, USA) was added to each reaction, and the mixture was incubated for 1 hour at 37 °C. Then, the reaction mixture was plated on an LB plate. After incubation of the plates overnight at 37°C, the numbers of surviving bacteria were counted. Serum bactericidal titers were defined as the reciprocal of the serum dilution that resulted in killing more than 50% of the bacteria compared to that achieved by the bacteria-complement-sera from control mice (inoculated with adjuvant only).

Vaccinated or control mice (adjuvant only) were injected intra-peritoneally (IP) with 1 × 10⁴ CFU of NTUH-K2044 or NTUH-A4528 per mouse one week after the 2^nd boost (5 mice per group). The mortality of the mice was tracked for 30 days. Survival was analyzed by Kaplan-Meier analysis with a log-rank test; a P value <0.05 was considered statistically significant. Statistical analysis was performed using Graphpad Prism 5.