In Planta Application and Efficacy of FireFighter-A Phage Cocktail for Combating Fire Blight
Article information
Abstract
Fire blight, caused by Erwinia amylovora, poses a significant threat to Rosaceae crops and has caused substantial damage to South Korea since its emergence in 2015. Traditional control methods, including antibiotic- and copper-based treatments, have shown limitations, underscoring the necessity for alternative solutions to enhance efficacy while addressing the concerns associated with antibiotics. This study evaluated the efficacy of FireFighter-A, a novel phage cocktail, in managing fire blight. Comprising four phages—Fifi318, Fifi451, pEa_27, and pEa_47—FireFighter-A demonstrates a broad host range that covers all recent isolates against E. amylovora and E. pyrifoliae, along with high stability under various conditions, including temperature, pH, and buffers. Trials with immature apple fruits, tissue culture rootstock plantlets, acclimated M26 rootstock plantlets, and blossoms have demonstrated that FireFighter-A provides superior biocontrol efficacy through synergistic effects, outperforming individual phages. Moreover, it significantly reduced fire blight symptoms and infection rates with a performance comparable to or exceeding that of streptomycin. These findings support the use of FireFighter-A as an effective and environment-friendly alternative for fire blight control.
Erwinia amylovora, a plant pathogen responsible for fire blight in the Rosaceae family, poses a significant threat to agricultural economies worldwide (Pedroncelli and Puopolo, 2023). Since its unwelcome debut in 2015 in a pear orchard in Anseong, South Korea, fireblight outbreaks have surged rapidly across the nation, resulting in extensive public control measures, including the burial of 348 orchards between 2015 and 2019 (Ham et al., 2020). Moreover, there are increasing reports of environmental plant hosts of E. amylovora in Korea, including Korean mountain ash and apricots, which further complicate the challenges posed by this pathogen (Lee et al., 2021; Lim et al., 2023). The presence of environmental hosts other than apples and pears makes it increasingly difficult to prevent pathogen spread. Conventional protocols based on antibiotics, and copper-based treatments have shown limited success and have promoted antibiotic resistance transmission, highlighting the urgent need for alternative approaches (Gusberti et al, 2015; Stockwell and Duffy, 2012).
Bacteriophages (phages) offer a promising alternative for combating bacterial infections that are difficult to treat with antibiotics. As phages are performed with high specificity, only the targeted pathogens decline while sparing beneficial microorganisms in the system (Koskella and Meaden, 2013). Abundant in nature, phages are safe for plants, animals, and humans, with no harm to non-bacterial cells, and their specificity reduces the risk of resistance development and spread (Ashelford et al., 2003; Dennehy and Abedon, 2021). This safety profile makes phages environmentally friendly and sustainable choices for disease management (Holtappels et al., 2021). Phages are highly recommended and widely chosen as biocontrol agents, particularly in the plant agriculture sector, where numerous commercial phage products are readily available (Biosca et al., 2024; Frampton et al., 2012).
Owing to the notorious reputation of Erwinia, a considerable number of phages have been isolated and characterized (Müller et al., 2011; Thompson et al., 2019). With the growing diversity of phages, a multitude of strategies for their application against fire blight caused by E. amylovora have emerged (Gayder et al., 2024). These strategies include evaluating the antibiotic potential of individual phages, combining multiple phages to actively address phage resistance, integrating phages with conventional antibiotics, extending their persistence in plants using nonpathogenic carrier bacteria, and employing chronic phages that can modulate virulence (Akremi et al., 2020; Ibrahim et al., 2024; Kim et al., 2022). This diverse arsenal of phage-based approaches offers a promising avenue for reducing antibiotic usage while effectively combating E. amylovora.
Here, we present an Erwinia phage cocktail FireFighter-A, which comprises four phages. The biocontrol potential of FireFighter-A was demonstrated in assays using immature fruits, plantlets (M26), potted seedlings, and flowers upon treatment with a phage cocktail, compared to antibiotics (streptomycin) or single phages. The compatibility, dose dependency, and synergy of FireFighter-A with antibiotics demonstrate the usefulness of FireFighter-A as a biocontrol agent against fire blight caused by E. amylovora.
Materials and Methods
Microorganisms and culture condition
The reference Erwinia amylovora strain TS3128, the only strain permitted for research use in South Korea, was employed in this study. Previously reported phages, Fifi044, Fifi67, Fifi106, Fifi318, Fifi451, pEa_SNUABM_27 (hereafter, pEa_27), pEa_SNUABM_31 (pEa_31), pEa_SNUABM_32 (pEa_32), and pEa_SNUABM_47 (pEa_47), which have demonstrated potential as anti-fire blight agents, were used to generate a phage cocktail for subsequent assays (Kim et al., 2020, 2022, 2024; Park et al., 2022). All microorganisms used in this study were cultured in tryptic soy broth (TSB; Becton Dickinson, Franklin Lakes, NJ, USA) with shaking at 150 rpm or sub-cultured on tryptic soy agar (Becton Dickinson) at 27°C (Kim et al., 2024). For the phage experiment, top agar (TSB supplemented with 0.4% bacteriological agar) was used.
Bacteriophage propagation and purification
Phage propagation and purification were performed as previously described (Jo et al., 2024). The filtered phage lysate (109–1010 plaque-forming units [PFU]/mL) was precipitated with NaCl (0.5 M) and PEG 8000 (5% w/v) at 4°C for 18 h. The solution was centrifuged to collect the phages, which were further purified by CsCl gradient ultracentrifugation (gradient densities of 1.7, 1.6, 1.5, 1.4, and 1.3 g/mL) at 50,000 rpm for 3 h. Phage bands (blue) were collected and dialyzed (10,000 molecular weight cut-off [MWCO]) in SM buffer. The purified phage (approximately 1010 PFU/mL) was stored at 4 °C for further use.
Bioinformatic analysis of phage genome
Libraries for whole-genome sequencing of the phage were prepared using the Illumina TruSeq Nano DNA Kit (Illumina, San Diego, CA, USA), and sequencing was conducted on an Illumina HiSeq 2500 system at Macrogen (Seoul, Korea). Raw reads were trimmed using Trimmomatic (v0.36), and the filtered data were assembled de novo using SPAdes (version 3.12). Open reading frames were predicted using Prokka (version 1.12b), GenMarkS, RAST, BLAST, and HHpred. tRNAs were identified using ARAGORN and tRNAscan-SE (version 2.0). A phylogenetic tree was constructed using Virus Classification and Tree Building Online Resources (VICTOR).
In vitro antibacterial assay for screening the phages
The phage with the best anti-E. amylovora effects from previous studies were used to screen the cocktail phage components. To screen the phage cocktail and examine its bactericidal effects, an in vitro antibacterial assay was conducted as previously described (Kim et al., 2022). A screening assay was conducted using fractional combinations of nine candidate phages recently reported to exhibit anti-E. amylovora activity: five from Dr. Roh (Fifi044, Fifi67, Fifi106, Fifi318, and Fifi451) and four from Prof. Park (pEa_27, pEa_31, pEa_32, and pEa_47). The phages were formulated into 4- or 5-phage cocktails (combined at multiplicity of infection [MOIs] of 0.1 and 10, with each phage included at an equal ratio and selected based on distant genomic relatedness) and co-cultured with E. amylovora at 105 colony-forming units (CFU)/mL. Combinations that completely inhibited bacterial growth were selected for the FireFighter-A phage cocktail. All tests were performed in triplicates.
Host range assay
The infective spectra of the four selected phages for FireFighter-A were determined as described previously (Park et al., 2022). Briefly, a purified phage solution was serially diluted (101–108 PFU/mL) and spotted on lawn of blight causing Erwinia strains (E. amylovora and E. pyrifoliae) that were recently isolated. The number of plaques was calculated to derive the efficiency of plating (EOP), which was subsequently standardized to an arbitrary index (AI). The normalization involved assigning EOP values of 1, 0.1, 0.01, 0.001, 0.0001, and 0 to the AI scores of 5, 4, 3, 2, 1, and 0, respectively. All tests were performed in triplicates.
Transmission electron microscopy of selected phages for FireFighter-A
Morphological analysis of the phages was performed using a transmission electron microscope (Talos L120C, 120 kV) with negative staining using 2% uranyl acetate (Kim et al., 2020). The mixture (1:1) of the purified phage and uranyl acetate was spotted onto formvar/carbon grids for 1 min. The solution was removed, washed with distilled water, air-dried, and observed under a microscope. Three virions were measured to analyze the dimensions of the phages.
Stability assay of selected phages for FireFighter-A
To examine the stability of phages against environmental factors, including pH and temperature, phages were incubated under each condition for 1 h. For pH stability tests, the phage stock was inoculated into 1 mL of SM buffer adjusted to pH 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, and 11.0, using 1 M NaOH or 1 M HCl to achieve a final concentration of 105 PFU/mL. The tubes were then incubated at 27 °C. For thermal stability tests, 1 mL phage suspension (105 PFU/mL) was incubated at 30, 40, 50, and 60°C and titrated for phage activity. The stability of the phages in the presence of ultraviolet (UV) light was tested using a large-area illumination source (100 mW/cm2, Oriel Sol1A Class ABB Solar Simulator, Newport, Yongin, Korea) that emitted light with a spectrum similar to that of the solar spectrum, as described previously (Jo et al., 2023). The phage solution was irradiated for 6 h and samples were collected every hour to confirm phage activity. Buffer stability tests of the phages were conducted at 27°C, with the initial phage titer adjusted to 108 PFU/mL by diluting phages in 1×, 1/100, 1/1,000, and 1/10,000 SM buffer, and the phage titer measured over 28 days. The phage activity was measured using the double-layer agar method. All tests were performed in triplicates.
Biocontrol assay in immature fruit
The biocontrol effect of the phage cocktail FireFighter-A was examined in the immature fruit of Fuji apples (Malus domestica), as previously described (Kim et al., 2024). The fruits were surface-sterilized using ethanol (70%), dried, pricked with a sterilized pipette tip, and pre-treated with each agent, followed by inoculation with E. amylovora. The agents included sterilized distilled water (control), each constituent phages of FireFighter-A, FireFighter-A, AgriPhage, and streptomycin. In the cocktail FireFighter-A, each phage was included at an equal ratio. The inoculum (10 μL) for bacteria, phages, and streptomycin was 105 CFU, 108 PFU, and 100 ppm per fruit, respectively. The symptoms were recorded for 2 weeks in a humidified chamber (27°C, approximately 90% of relative humidity [RH]) and scored on six scales by water soaking and browning of fruit; 0 = 0% for both water soaking and browning, 1 = 1–10% for both water soaking and browning, 2 = 11–25% of water soaking and 11–35% of browning, 3 = 26–50% of water soaking and 36–55% of browning, 4 = 51–75% of water soaking and 56–80% of browning, and 5 = 76–100% of water soaking and 81–100% of browning (Kim et al., 2024). The assay was performed in multiple biological replicates, ranging from n =37 to n = 54.
Biocontrol assay in planta
The biocontrol efficacy of the FireFighter-A phage cocktail was examined in M26 apple tissue culture rootstock plantlets (Malus domestica), acclimated M26 apple rootstock plantlets (Malus domestica), and Blossom (Malus prunifolia). The agents included sterilized distilled water (control) and each constituent phage of FireFighter-A, FireFighter-A, AgriPhage, and streptomycin for all planta assays. For the tissue culture plantlet assay, plants were cultured in Murashige and Skoog’s medium supplemented with 0.8% (w/v) agar and 3% (w/v) sucrose in a vessel (72 × 72 × 100 mm), as previously described (Kim et al., 2024). Three-month-old plantlets were pre-treated with each agent and then inoculated with E. amylovora for 2 h post-treatment. The inoculum (1 mL) for bacteria, phages, and streptomycin was 105 CFU/mL, 108 PFU/mL, and 100 ppm per vessel, respectively, and was sprayed to cover the entire surface of the plantlets. The symptoms were recorded for 2 weeks in a humidified chamber (27°C, approximately 90% of RH). The results were obtained in five biological replicates. For the acclimated rootstock assay, plantlets were pre-treated with each agent and then inoculated with E. amylovora for 2 h post-treatment. The inoculum (2 mL) for bacteria, phages, and streptomycin was 105 CFU/mL, 108 PFU/mL, and 100 ppm per plant, respectively, and was sprayed to cover the entire surface of the plantlets. The symptoms were recorded for 2 weeks in a humidified chamber (27°C, approximately 90% of RH). The results were obtained from seven biological replicates. For the flower assay, each agent was pre-treated at full agent, followed by inoculation with E. amylovora 2 h post-treatment. The inoculum (2 mL) of bacteria, phages, and streptomycin was 105 CFU/mL (for general bacterial burden) or 106 CFU/mL (for a high bacterial burden), 108 PFU/mL, and 100 ppm per fruit, respectively. The inoculums were sprayed to cover the entire surface of the flowers. The symptoms were recorded every 5 days in a humidified chamber (27°C, approximately 90% of RH) and the infected flower count was recorded.
Statistical analysis
All experiments were performed with independent biological replicates, and the number of replicates (n) for each treatment group is specified in the corresponding methods section. In some cases, n-values varied among groups due to differences in sample availability or experimental recovery. For in vitro antibacterial and buffer stability assays, mean values were compared using one-way ANOVA followed by the Holm-Sidak test. For immature fruit assays, median disease index values were compared using the Mann-Whitney rank sum test. For in planta assays, the incidence of infection was analyzed using the Fisher’s exact test. A P-value < 0.05 was considered statistically significant. All statistical analyses were performed using SigmaPlot v12.5 (Systat Software, San Jose, CA, USA).
Results
Screening of candidate bacteriophages for cocktail phage, FireFighter-A
Among the nine previously screened bacteriophages that exhibited a broad host range and strong bactericidal activity, combinations of 4 or 5 phages from different phylogenetic clusters were tested (Supplementary Fig. 1). As testing full factorial combinations is impossible, representative combinations including 4 or 5 phage cocktails, were examined (Supplementary Fig. 2). Among the 4 or 5 phage cocktail combinations, if 2 phages in the genus Loessnervirus (Fifi044, Fifi451, and pEa_27) and 1 phage Kolesnikvirus (Fifi106 and Fifi318) were included in the combinations, no bacteriophage insensitive mutants emerged (Supplementary Fig. 2). To minimize the complexity, four phage combinations was chosen and the combination “4ϕO,” comprised Fifi318 (apple orchard origin), Fifi451 (pear orchard origin, wider host range than Fifi044), pEa_27 (stream water near apple orchard), pEa_47 (apple orchard origin, wider host range than pEa_32). This combination, which completely inhibited the growth of E. amylovora at both high and low MOIs, was selected as a potential candidate for the phage cocktail FireFighter-A (Fig. 1). It comprises two Loessnervirus (Fifi451 and pEa_27), one Kolesnikvirus (Fifi318), and one Eneladusvirus (pEa_47), as indicated by the black arrow in Supplementary Fig. 1.
Inhibitory effect of Erwinia amylovora (TS3128) by FireFighter-A at a multiplicity of infection (MOI) of 0.1 (A) and 10 (B). Data represent mean ± standard deviation (SD) from three independent experiments (n = 3). Statistical analysis was performed using one-way ANOVA with Holm-Sidak’s test (P < 0.05).
Biological characteristics of the candidates of FireFighter-A
The host range of each phage—Fifi318, Fifi451, pEa_27, and pEa_47—could not individually cover all the recently isolated blight-causing strains of E. amylovora and E. pyrifoliae. However, when used in combination, these phages covered all pathogens, indicating a broad host range (Supplementary Fig. 3). All the phages in FireFighter-A had myovirus-like virions with contractile tails (Fig. 2A). As the two phages, Fifi451 and pEa_27, were classified in the same genus, Loessnervirus, the morphology of their plaques (11–13 mm) and virions (head: 65 nm/tail: 120 nm) was not significantly different. Fifi318 (Kolesnikvirus) and pEa_47 (Eneladusvirus) demonstrated plaque sizes of 2–3 mm with turbid centers, and clear 1 mm plaque, respectively. Fifi318 and pEa_47 have head diameters of 74 and 127 nm and tail lengths of 110 and 195 nm, respectively. Four phages of FireFighter-A were obtained from different species clusters (Supplementary Fig. 1) and did not encode antibiotic resistance or virulence genes (Kim et al., 2020, 2022, 2024; Park et al., 2022). General information on the selected phages is summarized in Table 1.
Biological characteristics of the phages in FireFighter-A. Morphological characteristics of the virions (A). Virion structure was observed by transmission electron microscope (120 kV). Scale bars = 100 nm. Stability of the bacteriophages on thermal (B), pH (C), ultraviolet (UV) (D), and buffer (E) stress. The assay was performed in triplicate (B–E). PFU, plaque-forming units. Data represent mean ± standard deviation (SD) from three independent experiments (n = 3). Statistical analysis was performed using one-way ANOVA with Holm-Sidak test (P < 0.05).
General features of the selected phages for FireFighter-A
All four phages are highly stable against thermal stress up to 50°C; however, no infective phages were detected after 1 h of incubation at 60°C (Fig. 2B). All four phages were highly stable at pH 4–10 (Fig. 2C). However, they showed varying susceptibilities at pH 11: Fifi451 and pEa_27 exhibited a slight decrease (less than 1 Log PFU/mL), pEa_47 demonstrated a moderate decrease (2–3 Log PFU/mL), and Fifi318 experiences a significant decrease (5 Log PFU/mL). Simulated solar irradiation (1 SUN, 100 mW/cm2) for an average daily sunlight duration of 6 h decreased the infectivity of the four phages slightly: Fifi318 (1 log PFU/mL), Fifi451 (1.44 Log PFU/mL), pEa_27 (1.27 Log PFU/mL), and pEa_47 (1.53 Log PFU/mL) (Fig. 2D). Buffer stability was observed at 1/100 to 1/10,000 dilutions in filtered tap water over 4 weeks. The results in Fig. 2E show no buffering effect for Fifi451 and pEa_27 after 672 h, a moderate buffering effect for pEa_47 after 96 h, and a strong buffering effect for Fifi318 after 24 h. The natural decline in concentration for Fifi451 and pEa_27 at room temperature was less than 1 Log PFU/mL for 4 weeks in SM buffer, 1/100 diluted SM buffer, 1/1000 diluted SM buffer, and 1/10,000 diluted SM buffer at 27°C. The Fifi318 concentration gradually decreased starting 24 h after dilution, and no infective virions were observed after 504 h in any of the diluted samples. Phage pEa_47 in the SM buffer (no dilution) showed a decrease of under 1 Log PFU/mL over 672 h. However, the decline in concentration for all dilutions began on day 4, with a total decrease of 2 Log PFU/mL by the end of the 672-h period.
Biocontrol effect of FireFighter-A in immature apple fruit
Immature apple fruits were wounded and pretreated with 10 μL of control agents (Streptomycin, AgriPhage, single phages, and FireFighter-A) followed by inoculation with E. amylovora. After a 7-day incubation period, the symptom index was evaluated based on water soaking and necrosis, which are typical symptoms of fire blight. The control group had a mean symptom score of 2.9, while streptomycin treatment completely inhibited the incidence of fire blight (Fig. 3). AgriPhage treatment reduced the progression of fire blight (symptom score, 2.4; P < 0.05). The effect of each single phages of FireFighter-A treatment varied, with symptom scores ranging from 0.7 to 3.1, and the cocktail FireFighter-A significantly reduced fire blight, with an average symptom score of 1.3, which was lower than that of AgriPhage (P < 0.001). Representative images of each treatment are shown in Fig. 3B.
Evaluation of anti-fire blight effect of phages in immature apple fruits. The inoculum was adjusted to 105 colony-forming units (CFU), 108 plaque-forming units (PFU), and 100 ppm per fruit for Erwinia amylovora (TS3128), phages, and streptomycin, respectively. (A) Box plot distributions of mean fire blight symptom index. The line in the box indicates the median, the colored star (★), and blank star (⋆) indicates the mean, and median, respectively. Vertical bars indicate the interquartile range between the 1st and 3rd quartiles. (B) The representative pictures that showing the result of each agent at the end point. Statistical analysis was performed using the Mann-Whitney rank sum test, and different letters indicate statistically significant differences (P < 0.05).
Biocontrol effect of FireFighter-A in planta
The in planta biocontrol effect of FireFighter-A on E. amylovora was evaluated using three models: tissue culture rootstock plantlets, acclimated M26 rootstock plantlets, and flowers. For tissue cultures of rootstock plantlets, the preventive effect was analyzed based on the rate of infected leaves (Fig. 4). The only agents that significantly prevented infection by E. amylovora were streptomycin (P = 0.01), each phage of FireFighter-A and FireFighter-A (P < 0.001) itself. The infected leaf rates of the positive control and AgriPhage were 62.4% and 52.0% (P = 0.15), respectively, whereas streptomycin (44.68%; P = 0.01) significantly prevented the infection based on the infected leaf rate as shown in Fig. 4B. Single phages and FireFighter-A enhanced the biocontrol effect even more, showing significantly higher efficacy than both streptomycin- and AgriPhage-treated groups. The acclimated M26 rootstock plantlet model was examined as a tissue culture rootstock model (Fig. 5). FireFighter-A demonstrated the most effective biocontrol effect on the number of infected leaves, with only 10.43% infected leaves (P < 0.001) (Fig. 5A). In comparison, streptomycin and AgriPhage infected 18.18% and 19.74% of leaves, respectively, and the representative images are shown in Fig. 5B. To examine the prevention of fire blight by FireFighter-A in blossoms, bacterial challenge was performed at two doses: 105 CFU/mL (low dose) and 106 CFU/mL (high dose). At the general bacterial burden E. amylovora challenge (Fig. 6A), all agents significantly prevented fire blight progression, with 4.68%, 5.16%, and 10.47% of flowers infected with streptomycin, AgriPhage, and FireFighter-A, respectively (positive control: 40.41%). In the high-dose E. amylovora challenge (Fig. 6B), the positive control showed 86.96% infected flowers, whereas the effectiveness of all agents decreased compared with the low dose of bacterial challenge. Streptomycin remained significant, showing 12.83% flower infection, which was comparable to the double treatment with FireFighter-A (11.48%). However, the preventive effect of AgriPhage was significant compared with that of both the positive control and other agents (streptomycin and single/double treatment with FireFighter-A). No phytotoxicity was observed following treatment with either a single phage or FireFighter-A.
The biocontrol effect of FireFighter-A in M9 apple rootstock plantlets. The inoculum was adjusted to 105 colony-forming units (CFU)/mL, 108 plaque-forming units (PFU)/mL, and 100 ppm for Erwinia amylovora (TS3128), phages, and streptomycin, respectively. The bar graph represents the ratio of infected leaves (A) and the representative pictures that showing the result of each agent at the end point (B). Statistical analysis was performed using Fisher’s exact test, and different letters indicate statistically significant differences (P < 0.05).
The biocontrol effect of FireFighter-A in acclimated M9 apple plantlets. The inoculum was adjusted to 105 colony-forming units (CFU)/mL, 108 plaque-forming units (PFU)/mL, and 100 ppm for Erwinia amylovora (TS3128), phages, and streptomycin, respectively. For the assessment of disease prevention, infected leaves were counted. The bar graph represents the ratio of infected leaves (A) and the representative pictures that showing the effect of each agent at the end point (B). Statistical analysis was performed using Fisher’s exact test, and different letters indicate statistically significant differences (P < 0.05).
The biocontrol effect of FireFighter-A in blossom. The inoculum was adjusted to 105 colony-forming units (CFU)/mL (A) and 106 CFU/mL (B) for Erwinia amylovora (TS3128), 108 plaque-forming units (PFU)/mL for phages, and 100 ppm for streptomycin. Statistical analysis was performed using Fisher’s exact test, and different letters indicate statistically significant differences (P < 0.05).
Discussion
E. amylovora is a devastating plant pathogen that has recently invaded apple and pear orchards in South Korea. While strict post-infection control measures, such as burial of infected trees, have been implemented, preventing the pathogen’s nationwide spread has proven difficult. From a sustainability perspective, phages offer a preferable alternative to antibiotics in controlling such pathogens, particularly to reduce the emergence of antibiotic resistance.
Lehman (2007) pioneered the development of phage-based biocontrol methods for fire blight, demonstrating significant efficacy of phages compared to antibiotics like streptomycin. Subsequent studies reported varying degrees of success, with phage treatments showing comparable effectiveness to antibiotics such as oxytetracycline and streptomycin, or biocontrol agents based on Bacillus species (Nagy et al., 2012). For example, Born et al. (2015) demonstrated the enhanced bactericidal effect of a two-phage cocktail comprising Erwinia phage Y2 (Loessenervirus) and M7 (Kolesnikvirus) compared to single-phage treatments in vitro as our observation in the present study (Fig. 1). Building on this foundation, other studies have explored combinatorial phage cocktails, including two-phage combinations tested on plant materials (Kim et al., 2024), three-phage cocktails (Hassan et al., 2023; Nagy et al., 2012), and cocktails containing four (Akremi et al., 2020), five (Kim et al., 2022), or even six phages (Biosca et al., 2024). These findings collectively underscore the potential of diverse and strategically combined phages to achieve superior biocontrol outcomes and reduce the risk of resistance emergence (Woods and Read, 2023).
Commercial phage products such as Erwiphage in Europe, AgriPhage in North America, and emerging products like Fire Quencher, have demonstrated mixed results. These products often rely on phage combinations to enhance efficacy, but limited diversity within their phages has, at times, hindered their ability to prevent resistance. For example, Erwiphage contains two similar siphoviruses (phiEaH1 and phiEaH2), while the first European phage cocktails include six genetically similar myoviruses from the Kolesnikvirus group, closely related to Fifi318 (Supplementary Fig. 1) (Biosca et al., 2024; Grace et al., 2021). In contrast, FireFighter-A was developed with greater genetic diversity (Supplementary Fig. 1), which strengthens its effectiveness and reduces the likelihood of resistant mutant emergence, even though each phage in the cocktail was included at only one-quarter of the concentration used in single-phage treatments (Supplementary Figs. 1 and 2). Notably, the optimal combination for enhanced anti-E. amylovora includes a diverse array of phages from different genera: Loessnervirus (Fifi044, Fifi451, or pEa_27), Kolesnikvirus (Fifi106 or Fifi318), and Eneladusvirus (pEa_47), as seen in cocktails such as 4ϕJ, 4ϕN, 4ϕO, and 5ϕF (Supplementary Fig. 2). The second-best combinations incorporated Alexandravirus (e.g., pEa_31 or pEa_32) in place of Eneladusvirus, adding another layer of genetic diversity. Conversely, combinations with limited diversity, such as 4ϕC (three Loessnervirus phages plus one Alexandravirus), failed to inhibit pathogen growth effectively, highlighting the limitations of low diversity for biocontrol. The precise synergistic mechanisms behind the superior performance of the four-phage cocktail, FireFighter-A, remain unclear and warrant further investigation. However, our findings suggest that there were no negative interactions, such as superinfection exclusion or competition for phage receptors, both of which could reduce cocktail efficacy (Hunter and Fusco, 2022). As demonstrated in this study, the phages within FireFighter-A do not exhibit exclusion of co-infection, enabling them to act synergistically (Abedon, 2015). Possible mechanisms for this synergy include sequential adsorption kinetics, and targeting of distinct bacterial receptors or metabolic states, which together may reduce the emergence of resistant mutants (Kim, 2025). However, we note that resistance emergence was only monitored over 24 h in this study, and longer-term evaluations are required to fully assess the durability of FireFighter-A. This underscores the importance of combining genetically diverse phages that do not compete with each other, as exemplified by FireFighter-A, to achieve effective biocontrol (Merabishvili et al., 2023).
These findings reinforce that phage cocktails, such as FireFighter-A, consistently outperform single-phage treatments in reducing fire blight severity in immature fruit assays. In our previous study, the single application of phages Fifi044 (Loessenervirus) and Fifi318 (Kolesnikvirus) demonstrated limited biocontrol efficacy against E. amylovora YKB14808 in immature fruit assays, yielding symptom scores of 2.27 and 2.81, respectively (Kim et al., 2024). In contrast, a two-phage cocktail combining these phages achieved a significantly improved symptom score of 1.81, highlighting the superior biocontrol effect of phage combinations. The four-phage cocktail FireFighter-A exhibited an even greater reduction in disease severity, with a symptom score of 1.3 against E. amylovora TS3128 (Fig. 3). This was significantly more effective than individual phage treatments (Choe et al., 2023; Kim et al., 2024). The combinatorial effects of phages have been well-documented in previous studies using immature fruits model. For example, a three-phage cocktail consisting of EAP1, EAP2, and EAP3 reduced disease severity to 12%, compared to 16–44% for individual phages (Hassan et al., 2023). Similarly, a four-phage cocktail containing Pear1, Pear2, Pear4, and Pear6 achieved a 40% reduction in symptoms, outperforming the individual phages (Akremi et al., 2020). Notably, the first European E. amylovora phage cocktail, which included vEam_PM_6, vEam_PM_21, vEam_S_24, vEam_W_25, vEam_PM_27, and vEam_W_28, demonstrated superior efficacy in immature fruit assays compared to single phage treatments (Biosca et al., 2024).
In planta validation of phages’ potential as biocontrol agents would confirm their efficacy in controlling disease under real-world conditions. FireFighter-A’s effectiveness was demonstrated through in planta assays using apple plantlets and young rootstocks (cv. Fuji) for shoot blight prevention (Figs. 4 and 5). At an MOI of 103, FireFighter-A consistently achieved greater disease control compared to individual phages, such as H5K or EA104, even when the latter were applied at an MOI of 108 (Nagy et al., 2015). Moreover, FireFighter-A’s individual phages outperformed single-phage treatments, such as Fifi106 which was documented in our previously reported (Choe et al., 2023). Flower is another critical pathway for E. amylovora dissemination and infection, making it essential to evaluate biocontrol efficacy in flower assays (Fig. 6) (Farkas et al., 2012). Previous studies have shown that effective fire blight control in flowers typically requires high phage densities ranging from MOI 105–108 (Nagy et al., 2015; Schwarczinger et al., 2017). However, additional strategies, such as employing phage carrier strains to enable in situ replication of phages, have demonstrated significant disease prevention with initial MOIs as low as 10 (Boulé et al., 2011). In contrast, FireFighter-A achieved significant disease prevention at much lower MOIs of 10–100, relying solely on the synergy of its phage combination, without the need for additional strategies. Its efficacy was further enhanced when additional treatments were applied (Fig. 6B).
In field trials, however, some reports indicate that phage-based approaches have not consistently achieved significant disease control (DuPont et al., 2023; Gdanetz et al., 2024; Koskella and Meaden, 2013; Koskella et al., 2022). One potential reason for these failures is the specificity of phages to their host E. amylovora strains (Holtappels et al., 2023). Our host range assay demonstrated the effectiveness of FireFighter-A phages against recent E. amylovora isolates from South Korea (Supplementary Fig. 3). FireFighter-A completely inhibited bacterial growth (Fig. 1), whereas the AgriPhage exhibited lower infectivity, resulting in limited cell lysis in vitro. This supports the common observation that strong host-phage interactions are crucial for effective biocontrol (data not shown; Lavelle et al., 2023). Similar results were obtained in immature fruit assays (Fig. 3), where FireFighter-A showed significant fire blight control, comparable to streptomycin, in planta tests using plantlets, potted seedlings, and flowers (Figs. 4–6). These findings underline the importance of selecting phages with broad host ranges and strong host specificity for successful biocontrol applications (Fong et al., 2021).
The other potential factors for failure of phage-based biocontrol have been suggested to environmental factors such as UV rays. A recent study by DuPont et al. (2023) highlighted the variability of phage efficacy compared to other treatments, including antibiotics. The authors identified UV radiation as a significant challenge for phage-based treatments, as over 99% of AgriPhage’s phages lost infectivity after 6 h of exposure to 21.6 MJ/m2 solar power. Although some protective compounds, like peptone and kaolin, have shown promise in lab settings (Born et al., 2015; Jo et al., 2023), their efficacy in field conditions remains to be fully validated. Employing adjuvants to shield phages from environmental stresses could potentially enhance the success of phage biocontrol. However, as noted by Gdanetz et al. (2024), other factors may play a more significant role than UV exposure, as only 2 out of 12 field tests using UV protectants demonstrated improved efficacy. Encouragingly, the four phages in FireFighter-A exhibited high stability under environmental stressors (Fig. 2), a characteristic shared with other phages reported by Akremi et al. (2020) and Gayder et al. (2024). In storage, FireFighter-A remained stable for 28 days in SM buffer at 27°C without direct sunlight, making it suitable for practical use in orchards without refrigeration. However, after dilution, it is recommended to use the solution immediately, as the concentration of the major constituent, Fifi318, begins to decline after 96 hours, with complete loss of infectivity occurring between 21 and 28 days (Fig. 2).
Phage-host interactions play a pivotal role in the success of biocontrol strategies. To date, numerous Erwinia phages have been reported for their fire blight control effects; however, our understanding of interaction between E. amylovora and its phages remains limited. This knowledge gap hinders the optimization of phage-based therapies and the development of strategies to counteract resistance. Resistance to phage Y2 (Loessenervirus), for instance, has been shown to affect the adsorption of distantly related phages, leading to cross-resistance (Knecht et al., 2022). This includes reduced infectivity or EOP for phages such as Bue1 (Ackermannviridae; Nezavisimistyvirus), L1 (Autographiviridae; Elunavirus), S2 (Autographiviridae; Eracentumvirus), and S6 (Schitoviridae; Waedenswilvirus), primarily due to modifications in the bacterial lipopolysaccharide (LPS) structure. However, M7, a member of Kolesnikvirus, maintained its infectivity against the resistant strain, suggesting it utilizes a receptor independent of LPS modifications. Our previous study demonstrated that resistance to jumbophages, such as pEa_SNUABM_31 and pEa_SNUABM_32 (Alexandravirus), pEa_SNUABM_47 (Eneladusvirus), and pEa_SNUABM_48 (unclassified singleton virus), resulted in cross-resistance among these phages (Kim et al., 2022). Despite this, pEa_27 (Loessenervirus) retained its ability to infect the resistant mutants with a maintained EOP, underscoring its potential as a critical component in phage cocktail formulations. These findings align with our results, which indicate that the most effective phage cocktail combinations include phages targeting exopolysaccharides, such as Loessenervirus, Nezavisimistyvirus, Eracentumvirus, and Waedenswilvirus, along with Kolesnikvirus and jumbophages. Ongoing studies are expected to elucidate the infection mechanisms of previously unexplored phages, such as Kolesnikvirus and jumbophages, further enhancing their application in biocontrol strategies (Nikolich and Filippov, 2020).
In summary, the genetic diversity and environmental stability of FireFighter-A make it a promising candidate for fire blight control in South Korea. It aligns well with existing orchard spray practices, particularly when used with adjuvants like kaolin or peptone to improve phage adherence and biocontrol efficacy (Jo et al., 2023). FireFighter-A could be deployed during critical infection windows, such as pre-bloom and early bloom, reducing the need for frequent antibiotic applications. As part of an integrated pest management strategy, FireFighter-A offers a sustainable, non-antibiotic alternative for fire blight management. Future field trials will aim to optimize treatment parameters under diverse environmental conditions, with early laboratory data suggesting up to a 90% reduction in infection rates and minimized resistant mutants compared to standalone phage treatments.
Despite its promising potential, further research is needed to explore the interaction of FireFighter-A with antibiotics and other control methods, as well as to investigate the mechanisms behind phage–bacteria dynamics. The current cocktail displayed encouraging efficacy; however, direct comparisons with previously reported combinations under identical experimental conditions were not conducted. Additional work is required to establish its relative performance, which will be critical for advancing phage-based fire blight control strategies in South Korea and beyond.
Notes
Conflicts of Interest
S.G. Kim, B. Kim, S.Song, S.C. Park, and E. Roh have applied for a patent application regarding this research (domestic patent application No. 10-2022-0073456).
Acknowledgments
This work was supported by the Rural Development Administration (RDA; grant number RS-2020-RD008879). M9 apple rootstock tissue culture plantlets were provided by the Korea Agriculture Technology Promotion Agency (Iksan, Republic of Korea).
Electronic Supplementary Material
Supplementary materials are available at The Plant Pathology Journal website (http://www.ppjonline.org/).