We purchased C. difficile ATCC 9689^T from ATCC (Manassas, VA, USA), and three clinical isolates of C. difficile (KPN10P03654, KPN10P03780, and KPN10P03783) from the Pathogen Resource Bank of Kyungpook National University Hospital (Daegu, Republic of Korea). The seven strains of C. difficile (CD-H-1, CD-H-2, CD-H-7, CD-M-5, CD-M-7, CD-M-9, and CD-M-11) used in this study were isolated from fecal samples of inflammatory bowel disease patients at Kyungpook National University Hospital in Daegu, Republic of Korea, between 2020 and 2022. All strains were cultured anaerobically inside an anaerobic gas chamber (BACTRON300; Sheldon Manufacturing Inc., Cornelius, OR, USA) at 37°C for 24-48 h in Mueller Hinton broth (MHB; BD, Franklin Lakes, NJ, USA), Mueller Hinton agar (MHA; BD, Franklin Lakes, NJ, USA), Brain Heart Infusion broth (BHI; BD, Franklin Lakes, NJ, USA), Brain Heart Infusion agar (BHA; BD, Franklin Lakes, NJ, USA), and blood agar plates (BAP; Synergy Innovation Co. Ltd., Gyeonggi-Do, Republic of Korea).

Escherichia coli BL21 (DE3) Star® was transformed with CHAP^SAP26-161 (483 bp) plasmids. Transformants were incubated in 1 L of lysogeny broth at 150 rpm and 37°C with ampicillin (150 μg/mL) until culture absorbance reached OD 0.5 at 600 nm. We added isopropyl β-d-1-thiogalactopyranoside to a final concentration of 0.1 mM and incubated the culture solution at 18°C for 16 h. Then, we cooled the culture medium on ice and centrifuged it at 6,000 rpm (at 4°C for 10 min) to obtain the cells. These cells were resuspended in 30 mL of lysis buffer (50 mM Tris-HCl [pH 8.0], 200 mM NaCl, and 100 μM ZnCl₂). Ultrasonication was repeated 10 times for 5 min to disrupt the bacterial cells and centrifuged at 15,000 rpm (at 4°C for 30 min). After centrifugation, the supernatant was filtered through a 0.22 μm syringe filter (GVS Abulo; Stanford, ME, USA). The filtered solution was then purified using a 5 mL His-trap column (GE Healthcare; Chicago, IL, USA) installed in a high-speed protein liquid chromatography system (GE AKTA Prime Plus FPLC System; Uppsala, Sweden). Residual imidazole was removed by dialysis against a lysis buffer of Tris-HCl (20 mM, pH 8.0) and NaCl (0.5 M). The purity and size of the protein CHAP^SAP26-161 (19.7 kDa) were confirmed by SDS-PAGE, and the protein concentration was measured using a nanodrop spectrophotometer (ND-2000; Thermo Scientific; Waltham, MA, USA). Polyclonal rabbit anti-mouse immunoglobulin G was conjugated with recombinant abysin and horseradish peroxidase, using an anti-His-Tag monoclonal mouse antibody (Ab Frontier, Republic of Korea) as the first antibody and 6X His-tag as a secondary antibody for Western blot analysis.

The MIC and MBC values of CHAP^SAP26-161 were determined using a broth microdilution method with a slight modification, as described by the Clinical and Laboratory Standards Institute (21), in round-bottomed 96-well microplates (SPL Life Sciences, Pochan, Gyeonggi-do, Republic of Korea). The exponentially grown bacterial cells were diluted to 10⁵ CFU/mL concentration in MHB and then incubated with CHAP^SAP26-161 (1-100 μg/mL) at 37°C for 24 h. We defined MIC as the lowest antibacterial concentration that completely inhibited bacterial growth. To determine the MBC, we spotted 10 μL of the mixture from each 96-well MIC test plate on MHA to evaluate bacterial viability. MBC was defined as the lowest concentration of antimicrobial agent with no growth on the agar plate. All assays were performed thrice. Moreover, we included growth and sterility controls, with LysSAP26 used as a positive control for MIC and MBC determination (22, 23).

Analysis of toxins A and B at the DNA level by PCR amplification

DNA samples for PCR amplifications were prepared using the boiling method. Briefly, 1-2 colonies of culture were taken from the agar plate and resuspended in 200 µL of RNase-free water (Welgene Inc., Gyeongsan, Republic of Korea) in a 1.5 mL Eppendorf tube. To settle the bacterial debris, we boiled the suspension was boiled for 7-10 min and centrifuged it at 10,000× g for 5 min. The supernatant containing the genomic DNA was then transferred to clean microcentrifuge tubes and used for PCR amplification. The remaining aliquots were stored at 20°C for future use. We used the tpi-specific primers (tpi-F [5′-AAA GAA GCT ACT AAG GGT ACA AA-3′] and tpi-R [5′-CAT AAT ATT GGG TCT ATT CCT AC-3′]) to detect the product of the 230 bp tpi gene specific to C. difficile. To detect toxins A and B, we incorporated toxin A-specific primer sets (tcdA-F [5′-AGA TTC CTA TAT TTA CAT GAC AAT AT-3′] and tcdA-R [5′-GTA TCA GGC ATA AAG TAA TAT ACT TT-3′]) and toxin B-specific primer sets (tcdB-F [5′-GGA AAA GAG AAT GGT TTT ATT AA-3′] and tcdB-R [5′-ATC TTT AGT TAT AAC TTT GAC ATC TTT-3′]) to amplify the 369 and 160 bp tcdA and tcdB genes, respectively (24). PCR was amplified in 50 μL of combined TaKaRa Ex Taq (0.25 μL [5 U/μL]), 10X Ex PCR buffer (Mg²⁺ plus, 5 μL), dNTP mixture (4 μL [2.5 mM each]), each primer (0.2 μM), and DNA template (1.5 μL), with the following cycling conditions: 95°C for 3 min; 40 cycles of 30 s at 95°C, 1 min at 50°C, 90 s at 72°C, and 10 min at 68°C. Additionally, genomic DNA was extracted using the QIAamp® DNA Mini Kit (Qiagen; Hilden, Germany) according to the manufacturer’s instructions. We designed the primers to obtain full-length tcdB genes of 7.1 kb and amplified them with forward and reversed primer sets tcdB_start (5′-ATA GTA AAG GAG AAA ATT TTA TG-3′) and tcdB_stop (5′-CTA TTC ACT AAT CAC TAA TTG AGC-3′). PCR was amplified in 50 μL of combined TaKaRa LA Taq (0.5 μL [5 U/μL]), 10X LA PCR buffer (Mg²⁺ free, 5 μL), MgCl₂ (5 μL [final 2.5 mM]), dNTP mixture (8 μL [2.5 mM each]), each primer (0.3 μM), and DNA template (2 μL), with the following cycling conditions: 94°C for 1 min, 32 cycles of 20 s at 98°C, 30 s at 47°C, 8 min at 68°C, and 10 min at 72°C. Similarly, to further confirm the tcdA gene, we designed the tcdA-specific forward primer NK3 [5′-GGA AGA AAA GAA CTT CTG GCT CAC TCA GGT-3′] and reverse primer NK2 [NK2: 5′-ACC AAT AGA AGA TTC AAT ATT AAG CTT-3′] to amplify a 252 bp fragment of the tcdA gene under the following thermocycler conditions: 94°C for 1 min, followed by 30 cycles of 20 s at 98°C, 30 s at 50°C, 1 min at 68°C, and 10 min at 72°C. All oligonucleotides were synthesized by Macrogen (Seoul, Republic of Korea). PCR amplifications were visualized by Gel Documentation System (Atto; Tokyo, Japan) after final electrophoresis on a 1% agarose gel stained with EcoDye (Solgent; Daejeon, Republic of Korea).

Toxins A and B from 11 C. difficile strains were identified by sandwich enzyme-linked immunosorbent assay (ELISA). In particular, we detected toxin A by using the EDI^TM Fecal C. difficile Toxin A ELISA Kit (Epitope Diagnostic Inc.; San Diego, CA, USA), and toxin B by using the Epilisa^TM Fecal C. difficile Toxin B ELISA Kit (Epitope Diagnostic Inc.; San Diego, CA, USA). All results were analyzed and interpreted according to the manufacturer’s instructions.

The A549 cell line’s (Sigma-Aldrich; St. Louis, MO, USA) viability was measured by 3-(4,5-methylthiazol-2-yl)-2,5-diphenyl- tetrazolium bromide (MTT) assay using an MTT cell viability assay kit (Amresco Inc.; Solon, OH, USA) according to the manufacturer’s instructions. Briefly, A549 cells (1 × 10⁵ cells/well) were seeded in a 24-well plate containing RPMI 1640 medium (Sigma-Aldrich; St. Louis, MO, USA) and kept overnight at 37°C in a CO₂ incubator for attachment. The next day, we replaced the medium with fresh RPMI 1640 medium with various CHAP^SAP26-161 concentrations (25-1,000 μg/mL, RPMI 1640 medium/well) and allowed the cells to grow for 24 h. After removing the medium, we washed the cells with phosphate-buffered saline (PBS), added 250 μL of MTT solution (0.5 mg/mL) to each well, further incubated the cells for 2 h, and then added 250 μL of solubilizing solution (90% isopropanol, 0.01% Triton X-100, and 0.01N HC5l79l). Color development after the reaction was measured at 570 nm using a VersaMax™ microplate reader (Molecular Devices, San Jose, CA, USA). Each experiment was performed in triplicate.

Female neutropenic C57BL/6 mice (5 weeks old, 16-19 g; Orient Bio, Seongnam, Gyeonggi-do, Republic of Korea) were housed (5 mice/cage) under controlled conditions (50% humidity and a 12 h:12 h light:dark cycle; specific pathogen-free) and fed a rat/mouse 18% 5L79 diet (PMI Nutrition International, Brentwood, MO, USA). Three different strains of C. difficile (C. difficile ATCC 9689^T, CD-M-5, and CD-H-7) were used for CDI. An antibiotic mixture of kanamycin (0.4 mg/mL; Sigma-Aldrich, USA), gentamicin (0.035 mg/mL; Sigma-Aldrich; St. Louis, MO, USA), colistin (850 U/mL; Sigma-Aldrich; St. Louis, MO, USA), metronidazole (0.215 mg/mL; Sigma-Aldrich, USA), and vancomycin (0.045 mg/mL; Sigma-Aldrich; St. Louis, MO, USA) was prepared in the mice’s drinking water. We deployed a uniform CDI murine model with a modification, as described by Chen et al. (25) and De Wolfe et al. (1). Briefly, mice were randomly grouped into healthy control (HC), CHAP^SAP26-161 control, CDI control, and CHAP^SAP26-161-treated CDI groups. CDI control or CHAP^SAP26-161-treated mice were allowed to drink the antibiotic cocktail for 5 days followed by sterile drinking water for 2 days. Mice received a single dose of clindamycin (20 mg/kg, Sigma-Aldrich; St. Louis, MO, USA) one day before C. difficile inoculation. On day 1, mice were challenged with 200 μL of C. difficile cells (5.0 × 10⁹ CFU) via oral gavage. After 3 h of C. difficile challenge, they received 150-300 μg of CHAP^SAP26-161 endolysin in 200 μL of Dulbecco’s phosphate-buffered saline (1X D-PBS, Welgene, Republic of Korea) intrarectally. All mice were monitored for 7 days. On day 8, they were given oral gavage of Clostridioides (5.0 × 10⁹ CFU) again in the same way as mentioned earlier, and after 3 h, they received CHAP^SAP26-161 endolysin intraperitoneally in the same amount (200 μL). Fig. 1 provides a graphical representation of the CDI model. The mice were divided into 14 different groups as described below. They were observed for disease signs, such as diarrhea, hunched posture, wet tail, and weight loss. The diarrhea was graded as 0, 1, and 2 for normal, loose, and liquid stools or soiled tail, respectively, as described by Tam et al. (26). In addition, we recorded their symptoms, survival, and weight daily. All mice were monitored for 14 days after the C. difficile challenge and euthanized by CO₂ asphyxiation using dry ice according to the American Veterinary Medical Association guidelines (https://www.avma.org/sites/default/files/2020-02/Guidelines-on- Euthanasia-2020.pdf). Fecal, colon, and tissue samples from the cecum and large intestines were collected for histopathological analysis. Moreover, we observed the mice daily for mortality and morbidity; those that were judged to be in a moribund state were euthanized, and their fecal, colon, and tissue samples were collected for histopathological examination.

The mice were divided into 14 groups as follows:

Group 1. The HC group was injected with 200 μL of the PBS alone (5 mice/group).

Group 2. The first CHAP^SAP26-161 control group was injected with 150 μg of CHAP^SAP26-161 alone (5 mice/group).

Group 3. The second CHAP^SAP26-161 control group was injected with 300 μg of CHAP^SAP26-161 alone (5 mice/group).

Group 4. The third CHAP^SAP26-161 control group was injected with 500 μg of CHAP^SAP26-161 alone (5 mice/group).

Group 5. Negative control group receiving antibiotics was injected with only 100 μL of PBS (5 mice/group).

Group 6. First CDI control group infected with C. difficile ATCC 9689^T (4 mice/group).

Group 7. First endolysin-treated group infected with C. difficile ATCC 9689^T was treated with 150 μg of CHAP^SAP26-161 after 3 h of C. difficile challenge (4 mice/group).

Group 8. First endolysin-treated group infected with C. difficile ATCC 9689^T was treated with 300 μg of CHAP^SAP26-161 after 3 h of C. difficile challenge (4 mice/group).

Group 9. Second CDI control group infected with CD-M-5 (3 mice/group).

Group 10. Second endolysin-treated group infected with CD-M-5 was treated with 150 μg of CHAP^SAP26-161 after 3 h of C. difficile challenge (3 mice/group).

Group 11. Second endolysin-treated group infected with CD-M-5 was treated with 300 μg of CHAP^SAP26-161 after 3 h of C. difficile challenge (3 mice/group).

Group 12. Third CDI control group infected with Clostridioides CD-H-7 (3 mice/group).

Group 13. Third endolysin-treated group infected with CD-H-7 was treated with 150 μg of CHAP^SAP26-161 after 3 h of C. difficile challenge (3 mice/group).

Group 14. The third endolysin-treated group was infected with CD-H-7 and treated with 300 μg of CHAP^SAP26-161 after 3 h of C. difficile challenge (3 mice/group).

The Animal Care Committee of Kyungpook National University (Daegu, Republic of Korea) approved this experiment (2022-0339-2).

Histopathological analyses were conducted to evaluate the inflammation and mucosal damage of the colon or cecum tissues caused by CDI. Cecum or colon tissue samples were fixed in 5% paraformaldehyde, embedded in paraffin, and sliced into 3 µm-thick sections. The deparaffinized sections were then stained with hematoxylin and eosin (H&E). We also stained them with periodic acid-Schiff (PAS) by using the PAS Stain Kit (Mucin Stain) (ab150680) (Abcam; Cambridge, United Kingdom) according to the manufacturer instructions. Using an Olympus SZX12 microscope (Tokyo, Japan), we observed and captured images of the tissue sections.

Bacterial DNA was extracted from fecal samples collected from euthanized mice, using a FastDNA® spin kit for soil (MP Biomedicals, CA, USA) according to the manufacturer’s instructions. The 16S rRNA genes were amplified using the universal primer pairs 338F (5′-ACT CCT ACG GGA GGC AGC AG-3′) and 806R (5′-GGA CTA CHV GGG TWT CTA AT-3′) corresponding to the V3-V4 region. These genes were sequenced using the Illumina MiSeq^TM PE300 platform at Accugene Inc. (Incheon, Republic of Korea). Briefly, the sequence data were processed with QIIME 2 (27). We filtered raw sequence data (FATSQ) and then removed chimeras by using CD-HIT-OTU and rDNA tools (28, 29). Subsequently, the assembled reads were clustered into operational taxonomic units (OTUs) using the UPARSE-OTU algorithm at a cutoff value of 97% identity at the species level (30). All representative sequences from each OTU were used for taxonomic assignment, and phylum-to-species level characterization was conducted using the Silva database (https://www.arb-silva.de). Alpha- diversity metrics (Shannon, Simpson, Chao 1, observed features, and Faith’s phylogenetic diversity (31)), beta-diversity metrics (weighted UniFrac, unweighted UniFrac, Jaccard distance, and Bray-Curtis dissimilarity), and principal coordinate analysis were estimated using q2-diversity. The differences between groups in taxonomic composition taxa were analyzed using the linear discriminant analysis (LDA) effect size (LEfSe) analysis, and LDA scores greater than 4 indicated discriminative taxa. Raw sequencing data have been uploaded to the NCBI Sequence Read Archive (SRA) database under BioProject ID: PRJNA1070869.

All statistical data were analyzed using OriginPro 2023b (OriginLab Corporation, Northampton, MA, USA). Kaplan-Meier survival analysis was used for the log-rank (Mantel-Cox) test. Statistical significance was measured by one-way ANOVA followed by Tukey’s test or Kruskal-Wallis test. Correlations between relevant variables were analyzed using Spearman’s rank correlation test. A P-value less than 0.05 was considered statistically significant.

The CHAP ^SAP26-161 gene was successfully cloned into the pET-21a (+) vector and expressed in E. coli cells bearing pET21a_CHAP^SAP26-161. The purified protein demonstrated the correct mass and near homogeneity with LysSAP26 (29.1 kDa) and CHAP^SAP26-161 (19.7 kDa) by SDS-PAGE analysis. The protein concentrations were 3.39 and 23.81 mg/L for LysSAP26 and CHAP ^SAP26-161, respectively (32).

The MIC and MBC values of CHAP^SAP26-161 against 11 strains of C. difficile were 50 µg/mL (Table 1). In all cases, these values had equal concentrations of LysSAP26 (75 µg/mL) and CHAP^SAP26-161 (50 µg/mL).

The species-specific internal fragment of the tpi gene was detected in the 11 strains of Clostridioides. Of the 11 isolates, eight toxigenic isolates (A+B+) showed tpi, tcdA, and tcdB amplification signals (in multiplex PCR as well), whereas three nontoxigenic isolates (A−B−) showed only tpi amplification signals using the primer sets tpi-F and tpi-R; tcdA-F and tcdA-R; and tcdB-F and tcdB-R. Interestingly, in immunogenic assays using ELISA kits, one isolate (CD-M-5) was found to be nontoxigenic (Table 2). Furthermore, we amplified the full size (7.1 kb) of the tcdB gene using primer sets tcdB_start and tcdB_stop and another flanking region of the tcdA gene using primer sets NK3 and NK2 to confirm the toxinotype. Finally, strain CD-M-5 was found to be nontoxigenic (A−B−).

A549 cells confirmed CHAP^SAP26-161’s cytotoxic ability in eukaryotic cells at specific concentrations, considering that epithelial cells represent the first line of defense against bacteria or bacterial products. The cell toxicity of the endolysin CHAP^SAP26-161 was evaluated on A549 human adenocarcinoma alveolar basal epithelial cells by MTT assay dose-dependently. CHAP^SAP26-161 showed no cytotoxic effect on the A549 cell line at 750 µg/mL or under, where the cell viability was more than 85% (cell viability > 80% are considered noncytotoxic) (33), indicating CHAP^SAP26-161 as a safe therapeutic agent. However, at 1,000 µg/mL, the cell viability was 74.8% ± 0.37%.

Change in body weight indicates the general health of mice during CDI. Body weight loss in CDI-induced mice is a marker of disease progression and represents CDI severity (25). All mice from the CDI group, which was infected with C. difficile spores of the strains ATCC 9689^T, CD-H-7, and CD-M-5, lost body weight compared with the HC group (Fig. 2a). Furthermore, all those from the HC group, negative control group, and second CDI control group infected with C. difficile CD-M-5 (A−B−) survived (Fig. 2b). Conversely, all mice from the first CDI control group infected with C. difficile ATCC 9689^T (A+B+) and the third CDI control group infected with CD-H-7 (A+B+) died after 8 (2 mice by day 6 and 2 mice by day 8) and 9 (one mouse by day six and two mice by day nine) days of postinfection, respectively (Fig. 2b). Importantly, all mice from the endolysin (CHAP^SAP26-161)-treated control groups infected with the C. difficile ATCC 9689^T and CD-H-7 survived (Fig. 2b). Oral gavage of CHAP^SAP26-161 (at a different concentration, 150 µg/mL and/or 300 µg/mL) improved not only body weight loss after infection but also diarrhea (Fig. 2a). In addition, CHAP^SAP26-161 recovered the colon lengths that were shortened because of CDI (Fig. 3). Immunohistological staining (H&E and PAS) showed that CHAP^SAP26-161 administration reduced histological injuries such as histological structure destruction and epithelial damage caused by CDI in the colon (Fig. 4).

Before endolysin therapy, CDI-induced significant alterations in gut microbiota composition. At the phylum level, the composition of the gut microbiota considerably differed between the CDI, HC, negative control, and endolysin-treated groups. The HC group mainly comprised Bacillota (92.4%), followed by Bacteroidota (5.9%), Verrucomicrobiota (0.5%), and Pseudomonadota (0.1%); the negative control group comprised Bacillota (89.1%), Bacteroidota (2%), Verrucomicrobiota (5.3%), and Pseudomonadota (3.6%); meanwhile, the CDI group showed a drastic reduction in Bacillota (44.3%) and an increase in Bacteroidota (2%), Verrucomicrobiota (5.3%), and Pseudomonadota (3.6%) (Fig. 5a). After endolysin treatment, the proportion of Bacillota increased to 60.5%, whereas those of Bacteroidota (19.7%) and Verrucomicrobiota (8.4%) decreased (Fig. 5a). Interestingly, the proportion of Pseudomonadota slightly increased to 11.2% (Fig. 5). At the family level, the proportions of Enterobacteriaceae (6.1%), Staphylococcaceae (3.2%), Porphyromonadaceae (6.2%), Peptostreptococcaceae (4.7%), Bacteroidaceae (29.2%), Erysipelotrichaceae (2.9%), and Christensenellaceae (0.7%) increased by 2,611-, 1,335-, 83.5-, 60.3-, 9.1-, 6.9-, and 3.2-fold, respectively, in the CDI group compared with those in the HC group (Fig. 5b). In addition, the proportions of 14 species, namely, Anaerostipes caccae, Bacteroides caccae, Blautia hansenii, C. difficile, Clostridium innocuum, Clostridium symbiosum, Enterocloster clostridioformis, E. coli, Limosilactobacillus reuteri, Mammaliicoccus sciuri, Parabacteroides goldsteinii, Phocaeicola vulgatus, Providencia alcalifaciens, and Thomasclavelia ramosa increased highly in the CDI group compared with those in the HC group (Fig. 5c). Importantly, the relative proportions of all these 14 species decreased in the endolysin treatment groups compared with those in the CDI group (Fig. 5c). In addition, the LEfSe analysis based on Linear discriminant analysis (LDA) score ≥ 4 showed that potential harmful bacteria such as C. difficile, E. coli, Mammaliicoccus sciuri which were abundant in the CDI group were much reduced in the endolysin treatment group. Thus, treatment with the endolysin CHAP^SAP26-161 may help restore the gut microbiota to a healthy balance by suppressing pathogenic microbes, including C. difficile and those associated with C. difficile. Moreover, endolysin treatment increased the proportions of Muribaculaceae, Akkermansiaceae, Lachnospiraceae, Ruminococcaceae, and Lactobacillaceae (Fig. 5b). Therefore, endolysin CHAP^SAP26-161 therapy could stabilize key microbial community structures associated with a healthy gut (Fig. 5).