All strains used in this study were obtained from the Phage Research Center of Liaocheng University, and the specific information is shown in Table 1 . The methods for the isolation of avian Salmonella are described in previously published articles (Cao et al., 2022), whereas S. Abortusequi of donkey origin was isolated into the lungs, hearts, livers, and spleens of donkeys suspected of having S. Abortusequi in various donkey farms in Shandong Province. Briefly, the lung, heart, liver and other tissues of aborted donkey foals were aseptically taken and inoculated in LB liquid medium for bacterial enrichment, and then streaked and cultured on Nutrient Agar and Salmonella Shigella (SS) Agar, respectively, and a single suspected colony was picked for purification culture, and then the purified strains were subjected to polymerase chain (PCR) and serotype identification (Wang et al., 2023).

Phage was isolated from the bedding of aborted donkeys. Briefly, 200 μL of S. Abortusequi stored at -80°C is placed in a 10 mL centrifuge tube and then 5-7 mL of liquid LB medium is added and shaken for 12-16 h at 37°C at 180 rpm. The bacteria were then purified on SS medium by three-zone delineation method, and a single colony was picked and inoculated in LB liquid medium at 37°C with shaking at 180 rpm for 12 h. After incubation, the culture solution was placed in a 4°C refrigerator for further use. An appropriate amount of bedding material was sampled in a 50 mL centrifuge tube, mixed well with the appropriate amount of sterilized saline, and centrifuged at 10,000 rpm for 5 min at 4°C to collect the supernatant, which was then filtered through 0.22 µm filter membrane. Then 1 mL of the filtrate and 200 µL of bacterial solution were added in a 10 mL centrifuge tube, mixed with 5-7 mL of liquid LB medium for 4 h at 37°C and 180 rpm shaking. The culture was then centrifuged at 10,000 rpm for 5 min at 4°C, and the supernatant was filtered through a 0.22 µm filter membrane. Finally, the phage was verified by the dot-drop method, and the presence or absence of phage spots was observed (Lu et al., 2003).

Host bacteria (200 µL) was mixed well with 6-8 mL of LB semi-solid medium melted, cooled to 55°C, poured into the plate, mixed with 5 µL of filtrate dropwise and incubated at 37°C for 4-6 h. Then, the phage plaques were picked up in centrifuge tubes containing 1 mL of SM buffer, mixed well using vibration, and then centrifuged to filter. The double-layer plate method was employed to purify the phage in the filtrate. The above steps of phage plaques deduction, centrifugation, and filtration were repeated 3-5 times until phage plaques of uniform size and morphology were obtained.

The double-layer plate method was used to prepare plates. Briefly, 10-15 mL of nutrient agar medium was melted and cooled to 55°C. The purified phage was diluted 10-fold, and the appropriate dilution was selected. Then, phage lysate (200 µL) and host bacteria (200 µL) were added in a 10 mL centrifuge tube, mixed well with 5-8 mL of LB semi-solid medium (melted and cooled to 55°C), and then poured into the configured plates placed in a 37°C incubator and cultured for 4-6 h. For each dilution, the experiments were repeated thrice.

The host range of isolated and purified phage was assessed by spotting 19 avian Salmonella strains with different serotypes and 7 S. Abortusequi strains of donkey origin, respectively (Cao et al., 2022). The phage host range assay is mainly performed using the double-layer plate spotting method. Briefly, for plate preparation, 200 µL of bacterial solution was mixed with 5-8 mL of melted LB semi-solid medium cooled to 55°C. To observe the presence or absence of phage plaques, 10 µL of phage lysate was dropwise added to the plates and incubated at 37°C for 4-6 h. The presence of phage plaques indicated that the phage could lyse the strain. The experiment was repeated three times to obtain reliable results.

For this analysis, the previously described method was followed (Cao et al., 2022). The phage was concentrated by double plate amplification to obtain highly efficient phage lysates for easy morphological analysis using a microscope. The phage lysate was negatively stained with 1% phosphotungstic acid, and the phage filtrate was added dropwise onto a copper mesh grid and then the morphology of the phage was observed using a JEM-1200EXII Transmission Electron Microscope (TEM) (JEOL, Japan) by imaging the isolated phage at an acceleration of 80 KV. The phage morphology was analyzed and measured using ImageJ software (Li et al., 2020b).

Using the 10-fold dilution method, the phage was diluted, and the appropriate dilution was mixed with host bacteria in a 1:1 ratio (500 µL each) at a concentration of 10⁸ CFU/mL at log₁₀ growth phase (1, 0.1, 0.01 and 0.001 respectively) for 3 h at 37°C with 180 rpm shaking. Then, the phage potency was determined using a double-layer agar plate method, and the optimal MOI with the highest potency was selected. For each group, 3 parallel experiments were conducted.

A one-step growth curve assay was performed with slight modifications of the previously described method (Sun et al., 2012) to assess the incubation period and phage outbreak size. The phage was diluted by the 10-fold dilution method, and the appropriate dilution was mixed with host bacteria in a 1:1 ratio (500 µL each) at a concentration of 10⁸ CFU/mL at log₁₀ growth phase and incubated in a water bath at 37°C for 10 min. This mixture was then centrifuged at 10,000 rpm for 5 min at 4°C to remove the supernatant, washed twice with equal volume of 37°C pre-warmed LB medium, resuspended with 1 mL of LB medium, added with 10 mL of LB medium, and incubated at 37°C with constant shaking at 180 rpm. The mixture was sampled (500 µL) at 0 min and then once every 10 min. Samples were taken at intervals of 120 min, and phage titers were determined immediately in the first portion, and in the other portion after treatment with 1% chloroform for 30 min, and then phage titers were determined, and the test was repeated thrice for each time point. The steps for the determination of phage titer were as follows (Zhao et al., 2024). The samples were centrifuged at 10,000 rpm for 3 min, filtered through a 0.22 µm filter membrane to obtain the phage lysate, and then diluted with a 10-fold specific dilution method. The titer of phage in the lysate was determined using the double-layer plate method. The one-step growth curve of phage was plotted with incubation time on the horizontal axis and the phage potency logarithmic value on the vertical axis. Phage burst size was evaluated by dividing the average number of final free phage particles by the number of initial free phage particles (Shang et al., 2021).

A total of 5 tubes of phage lysate (500 µL/tube) were incubated in a water bath at 40°C, 50°C, 60°C, 70°C, and 80°C, respectively. At each temperature, 100 µL of phage lysate was removed from each tube at 20, 40, and 60 min, respectively. The potency was assessed via the double-layer plate method. The experiment was repeated thrice.

Phage lysate (100 µL) was mixed with SM buffer (900 µL) at pH 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, and 13, respectively, in a water bath at 37°C for 1 h. The potency of the lysate was determined by the double-layer plate method. The experiment was repeated thrice.

For phage DNA extraction, E.Z.N.A^® DNA kit (OMEGA, United States) was employed, whereas TruSeq™ Nano DNA Sample Prep Kit (Illumina, United States) was used for sequencing libraries. Genome sequencing was performed by Shanghai Lingen Biotechnology (Shanghai, China) using the Illumina novaseq6000 sequencing platform. Quality control raw data were filtered using fastp 0.23.4 (Chen et al., 2018), and FastQC 0.12.0 (Brown et al., 2017), and the genome was assembled using SPAdes 3.15.5 (Bankevich et al., 2012).

The open reading frames of phage were predicted using GeneMarkS 4.28 (Besemer et al., 2001) (http://topaz.gatech.edu/GeneMark/genemarks.cgi). Furthermore, the function of each ORF-encoded protein was predicted via the BLASTp (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins ). Moreover, VFDB (Liu et al., 2019a) (http://www.mgc.ac.cn/cgi-bin/VFs/v5/main.cgi), CARD (Alcock et al., 2023) (https://card.mcmaster.ca/), ICEfinder (Liu et al., 2019b) (https://bioinfo-mml.sjtu.edu.cn/) were used, respectively, to predict the function of each ORF. Moreover, ICEfinder/ICEfinder.html predicted the genomic virulence genes, drug-resistance genes, and integrases. The ANI between genomes was calculated using JSpeciesWS (http://jspecies.ribohost.com/jspeciesws) (Richter et al., 2016) based on the Blast+ algorithm. The TBtools software was employed to plot ANI maps with ANI values. Phage LDDK01 has been uploaded to NCBI’s Genbank database under accession number PP932687.

The ability of phage LDDK01 to remove the host bacterial biofilm was determined by following the previous literature (Runci et al., 2017). Briefly, host bacterial solution (10⁸ CFU/mL, 200 µL) was added to a 96-well plate for 48 h at 37°C. Then, the culture solution was gently aspirated, washed with PBS buffer thrice, and incubated with phage LDDK01 lysate (200 µL) at 1, 0.1, 0.01, and 0.001 MOI. The control group received an equal volume of LB. The plates were incubated in a constant temperature incubator for 6 h and 12 h. Then, the wells were washed with 200 µL of PBS buffer, stained with 200 µL of 1% crystal violet solution at room temperature for 30 min, washed to remove the crystal violet solution thrice with sterile PBS buffer, and treated with 200 µL of 100% ethanol solution for 15 min to completely dissolved the biofilm. Lastly, 100 µL of this solution was sampled to measure the optical density (OD) at 595 nm with an enzyme meter.

The in vitro lysis ability of LDDK01 was determined using the reference already published in the literature (Lu et al., 2003). Briefly, different concentrations of phage (MOI = 1, 0.1, 0.01, and 0.001) were mixed with the host bacterial solution (10⁸ CFU/mL), respectively. The positive control group was the host bacterial culture medium without phage, while the negative control group included phage and LB medium; both the groups were incubated at 37°C for 24 h. The culture solution of each group was collected at 2-h intervals to measure the OD of each group at 600 nm. Each experiment was repeated three times.

The bacteriostatic effect of LDDK01 on S. Abortusequi present on the surface of donkey meat was verified based on the relevant methods described in the literature (Shang et al., 2021; Noor Mohammadi et al., 2022; Tian et al., 2022). Freshly slaughtered donkey meat was purchased from the slaughterhouse in Liaocheng City, Shandong Province. To reduce contamination, the meat (approximately 0.5 g) was sliced into 1 × 1 cm squares in a sterile biosafety cabinet, placed in a sterile culture dish, and disinfected with ultraviolet radiation. During the sterilization process, the petri dish was turned over twice, etc. After sterilization, donkey meat from different areas was randomly taken for bacterial isolation and identification, and phage LDDK01 biocontrol experiments were carried out after detection of sterility. Briefly, 50 μL of S. Abortusequi (10⁸ CFU/mL) was added to 5 spots (10 μL/spot) on the surface of each sterile donkey meat and incubated for 1 h in a sterile biosafety cabinet. Then, 100 μL of LDDK01 at 1, 0.1, 0.01, and 0.001 MOI were taken, added to the same location, and immediately placed in sterile sealed bags at 4°C and 25°C for 0, 1, 12, 24, 36, 48, and 72 h, respectively. For the positive control group, the same volume of sterile PBS was used instead of phages. At each sampling time point, donkey meat pieces were placed in 2 mL of sterile PBS buffer, minced with a tissue homogenizer, and then evenly divided into two parts. One part was centrifuged and diluted with PBS gradient for bacterial counting using SS medium at 37°C. The other part was centrifuged and filtered through a membrane to obtain a phage solution. The phage potency was determined at 37°C using a double-layer plate method. Each experiment was repeated thrice.

Bacterial counts and phage titers were determined by plate counting and double-layer plate methods, respectively. Furthermore, bacterial counts were converted to log₁₀ CFU/mL or log₁₀ CFU/piece, while the phage counts were converted to log₁₀ PFU/mL. All statistical analyses were performed using the SPSS 23.0 software (SPSS Inc., version 27.0; Chicago, IL, USA), and the data are expressed as the mean ± standard deviation (SD) of at least three biological replicates. Statistical differences between groups were assessed using Student’s t-test. p-values < 0.05 indicated a statistically significant difference (^ns p > 0.05, * p < 0.05, ** p < 0.01, and *** p < 0.001).

This study isolated an S. Abortusequi phage, named vB_SalP_LDDK01, from the bedding of a donkey farm in Liaocheng City, Shandong Province, China. The phage was purified using the double-layer plate method and then used to verify the host spectrum of 7 S. Abortusequi strains and 19 avian Salmonella strains with different serotypes ( Table 1 ). The results showed that phage LDDK01 lysed the 7 strains of S. Abortusequi, 2 strains of Salmonella typhimurium, and 4 strains of Salmonella enteritidis ( Table 1 ; Supplementary Figure S1 ). Furthermore, it formed transparent, regular, and uniformly sized round phage plaques on double-layer plates ( Figure 1A ). LDDK01 is a phage that has not been reported by other researchers. Therefore, it was further investigated. The TEM analysis of the purified LDDK01 revealed that it had an icosahedral head with a diameter of 54.8 nm and a long, non-contractile tail with a diameter of 120.6 nm ( Figure 1B ), suggesting that the phage belongs to the Siphoviridae family.

Different concentrations of LDDK01 and host bacteria were mixed and incubated for 3 h to determine the phage titers by the double-layer plate method ( Figure 2A ). It was observed that the phage potency was the highest at the MOI of 0.1, which was about 1.86 × 10¹⁰ PFU/mL; therefore, the optimal MOI of this phage was 0.1 ( Figure 2A ). One-step growth curve analysis showed that the incubation period of the phage was about 10 min. Furthermore, after the phage infected the host bacteria, its potency increased rapidly in 10-50 min and stabilized after 60 min ( Figure 2B ). The burst size of LDDK01 was calculated to be about 378 PFU/cell.

The heat resistance analysis revealed a stable phage potency at 40, 50, and 60°C for 1 h. After incubation at 70°C for 20 min, 40 min, and 1 h, the phage potency decreased by about 1.2 log, 1.7 log, and 2 log₁₀ PFU/mL, respectively. Moreover, its activity was completely lost at 80°C after only 20 min ( Figure 2C ). Similarly, LDDK01 maintained a high potency at pH 4-12; however, its potency was significantly reduced from pH 2 and 3 to 3 and 6.4 log₁₀₁₀ PFU/mL, respectively. At pH 1 and 13, the phage is completely inactivated under strong acid and alkali conditions ( Figure 2D ).

The whole genome sequence analysis revealed that LDDK01 had a 42,378 bp long genome with double-stranded DNA, 49.52% GC content, and encoded 57 ORFs. Furthermore, the annotation results showed that the ORFs of LDDK01 had 28 known proteins and 29 hypothesized proteins. The phage lacked tRNAs and did not carry virulence and drug-resistance genes. The predicted functional proteins primarily included DNA replication and modification modules (DNA ligase, endonuclease, etc.) and phage structural modules (tail protein, major coat protein, etc.). The annotation results of known functions of the phage genome are shown in Supplementary Table S1 , and the genome circle diagram is shown in Figure 3A .

For the isolated phage LDDK01, a proteomic tree was established using 5633 phage genomes acquired from the Viptree database. It was observed that LDDK01 and vB_SenS-EnJE1 were clustered together with a close affinity, and their bacterial hosts all belonged to the class Pseudomonadota ( Figures 3B, C ). The BLASTn comparison of the NCBI database revealed that the highest sequence homology (99.94%) was between the genome of LDDK01 and Salmonella phage vB_SenS_4FS1 (PP537413.1), followed by Salmonella phage PIZ SAE- 01E2 (MN336266.1) with 94.78% homology. The similar phage genome sequences were extracted and further analyzed using VIRIDIC ( Figure 3D ), which revealed that LDDK01 indicated the highest homology with Salmonella phage JD02 (OL652658), followed by Salmonella phage vB_SenS_Psm105 (PP437544) and Salmonella phage vB_SenS_Psm140 (PP437543).

The phylogenetic relationships between LDDK01 and 19 Salmonella phages were assessed based on two representative highly conserved proteins: the major coat protein (LDDK01-ORF12) and the terminase large subunit (LDDK01-ORF19). LDDK01 indicated a significant association with three phages of the genus Jerseyvirus, located in the same branch and distantly related to phages of eight other genera ( Figure 3E ). Furthermore, a similar evolutionary tree was obtained when the amino acid sequence of the terminase large subunit was employed ( Figure 3F ), which indicated a close association with two strains of Jerseyvirus and a distant affinity to Salmonella phage Jersey. Moreover, the genome sequences of LDDK01 were compared with those of the homologous phages vB_SenS-EnJE1 and JD02, respectively, using Easyfig to analyze the genome evolutionary relationships ( Figure 3G ). The data indicated that LDDK01 was significantly correlated with vB_SenS-EnJE1, showing similarities in the protein’s function and the regions where they were located. LDDK01 and JD02 showed opposite regional similarities. Overall, based on the morphology assessed via electron microscopy, comparative genomic analysis, phylogenetic analysis, and the latest guidelines of the International Committee on Classification of Viruses (ICTV), LDDK01 was determined as a new phage belonging to the Class Caudoviricetes and genus Jerseyvirus.

To verify the ability of LDDK01 to remove the biofilm of S. Abortusequi, the bacterial biofilm was treated with phage treatment and then subjected to the biofilm crystallization violet method. The results showed that only at MOI = 1, 6 h of phage treatment could remove the biofilm ( Figure 4A ). At 12 h of phage treatment, the phage significantly removed the biofilm ( Figure 4B ). These results suggest that LDDK01 can remove the S. Abortusequi biofilm. Furthermore, the in vitro inhibition of host bacteria within 24 h after LDDK01 treatment at different MOIs indicated that LDDK01 had good inhibitory effects in different infection complex groups compared with the bacterial fluid-positive control group. Moreover, the inhibitory effect of LDDK01 was similar at different MOI groups from 0 to 8 h, where the inhibitory effect at MOI = 1 was better than at other MOI after 8 h ( Figure 4C ).

This study investigated the inhibitory effect of LDDK01 on S. Abortusequi present on the surface of donkey meat. The sterile donkey meat pieces were artificially infected with S. Abortusequi and then treated with LDDK01 with different MOIs for different times at 25°C and 4°C, respectively. The results showed that LDDK01 significantly inhibited S. Abortusequi in donkey meat at different temperatures, and the bacterial inhibition rate was essentially proportional to the phage’s titer. The bacterial content of each group 4.3, 3.7, 3.3, and 3.5 log₁₀ CFU/piece after 72 h of LDDK01 treatment at different MOIs (1, 0.1, 0.01, 0.001) was compared with that of levels of S. Abortusequi in donkey meat from initial inoculation at 25°C ( Figure 5A ). Compared with the levels of S. Abortusequi in donkey meat from initial inoculation at 4°C, the bacterial content of each group decreased by 4.5, 3.9, 2.8, and 2.7 log₁₀ CFU/piece, respectively, after 72 h of LDDK01 treatment at different MOI ( Figure 5B ). These data suggested that LDDK01 is more effective in controlling Salmonella populations at 4°C than at 25°C. The phage assay ( Figures 5C, D ) showed that during the inhibition process at 25°C and 4°C, each group’s phage titers did not change significantly and were comparable to the initial titers.