Streptomycin-degradation tests were conducted using both hydrolysis and photolysis. In the hydrolysis test, streptomycin solutions were prepared in 50 mL conical tubes containing each 50 mg/mL of laboratory streptomycin (Streptomycin Sulfate, Fujifilm Wako Pure Chemical Co., Ltd. Osaka, Japan) and 250, 500, 1,000, and 2,500 μg/mL of bactericidal pesticide with streptomycin as 20% of active gradient (Agrepto, Kyungnong Co., Ltd, Seoul, Republic of Korea) and maintained at 28°C for 7 days in hydrolysis test. After 7 days, 10 μL of each streptomycin solution at 7 days was spot-inoculated on the center of mannitol glutamate yeast extract (MGY) medium that had been pre-plated with 100 μL of E. amylovora TS3128 bacterial suspension (Kang et al., 2021) at an optical density (OD) at 600 nm of 0.1 (approximately 3.1 × 10⁸ CFU/mL) and incubated at 28°C for 2 days. The effect of streptomycin degradation via hydrolysis was considered as an inhibitory zone against TS3128 by comparing the reduced size of the inhibitory zone with day 0 solution. In the photolysis test, 50 and 500 μg/mL of laboratory and bactericidal pesticide streptomycin solutions in 50 mL conical tubes, respectively, were exposed to an OD of 302 nm (UVB) using Ultraviolet Transilluminator (MaestroGen Inc., Hsinchu City, Taiwan) for 2 days, and TS3128 was then inoculated at a concentration of OD_{600 nm} = 0.3. The effect of UVB on streptomycin degradation was assessed by counting the colony forming unit (CFU/mL) of colonies grown on MGY medium without streptomycin at daily intervals until 5 days. All experiments were performed twice.

To confirm whether E. amylovora TS3128 cells entered the persistent state, which represents the biphasic kill curve after streptomycin treatment, each single colony of TS3128 was cultured in King’s B broth at 28°C for 20 h and then sub-cultured in the same liquid medium at 28°C for 5, 20, and 96 h to correspond to lag, early stationary, and late stationary phase, respectively. At the sub-culture phase, 50 and 500 μg/mL of laboratory and pesticide streptomycin were added to each phase-culture, respectively, and was shaken at 28°C until 5 and 8 h for lag and early and late stationary phases, respectively. To remove residents of streptomycin, 1 mL of samples were collected twice an hour, re-suspended in 10 mM MgCl₂ twice, and colonies were counted after 8 h on MGY agar medium. To date, a streptomycin resistant isolate of E. amylovora has not been found in Republic of Korea; therefore, a streptomycin resistant isolate of Pseudomonas syringae pv. actinidiae biovar 2, which was isolated in Republic of Korea, was used as a negative control (Psa2, Lee et al., 2020). Psa2 was cultured in King’s B broth with or without streptomycin for 20 h, and 1 mL of samples were collected and re-suspended with same procedures for TS3128. Colonies were then counted after 0, 12, and 24 h on King’s B medium with or without supplemented laboratory streptomycin (50 μg/mL).

To determine the end point of the persistence state induced by streptomycin, sub-cultivation was conducted in King’s B broth and distilled water, respectively. The CFUs of growing colonies were counted several times up to 72 h following previously described procedures. The experiments were conducted two times.

To prepare the early stationary phase during staining, E. amylovora cells were cultivated in King’s B broth at 28°C for 20 h as the nutrient rich case. The cultures were then centrifuged at 4,000 rpm for 6 min, re-suspended in 10 mM MgCl₂ twice, and the pellets were suspended in same volume of distilled water as lack of nutrient case. Following the preparation of the two case samples, 500 μg/mL of pesticide streptomycin was added to two samples, and 1 mL of samples were taken at 3 days intervals till 9 and 15 days for nutrient rich and deficient status, respectively, and rinsed in 10 mM MgCl₂ twice to remove streptomycin residue.

E. amylovora cells were stained with SYTO9 and propidium iodide (PI) using a BacLight Live/Dead viability kit (Invitrogen, CA, USA) to count total, live, and culturable cells according to previous reports (Santander et al., 2014; Kim et al., 2023). This was performed to produce a biphasic kill curve within the first few hours of cultivation, which showed a declining growth curve and identified the VBNC state after streptomycin treatment. To determine the proportions of each different physiological state in VBNC, staining with RedoxSensor Green (RSG) and PI was also conducted using the BacLight RSG Green vitality kit (Invitrogen, CA, USA) in accordance with a recommended protocol and previous report (Patel et al., 2021). Each 1 mL of the two case samples contained 1 μL of SYTO9 (3.34 μM) and PI (20 mM) or each of 1 μL RSG (1 mM) and PI and were incubated for 15 min in the dark. Subsequently, 200 μL of each stained samples were added to a FACS test tube and exposed for 1 min to FACSCalibur (BD Biosciences) and FACSymphony (BD Biosciences) for SYTO9 and PI and RSG and PI, respectively. FL1 530/30 bandpass filter for SYTO9 and FL3 670LP filter for PI in FACSCalibur and FITC (fluorescein isothiocyanate) 530/30 filter for RSG and PE 575/30 filter for PI in FACSymphony were used to capture fluorescence. Data analysis was conducted using the BD CellQuest Pro and BD FACSDiva software for Live/Dead and RSG kit, respectively. A total of nine fields were captured for fluorescence in three repeats. Four locations were identified among the recorded events (Supplementary Figure 1): PI-red (Q1), RSG and PI merge-yellow (Q2), no signal (Q3), and RSG-green signal (Q4). These zones were designated by 70% ethanol-killing, persistence and VBNC state, unstaining, and live cells, respectively. The CFUs of the culturable cells grown on MGY medium were counted as culturable cells in both staining.

To find alternative reagents that can shorten the duration of the persistence and VBNC state of E. amylovora cells induced by streptomycin, early stationary phase cells in distilled water were treated with 500 μg/mL of pesticide streptomycin and 10 mM of carbon sources such as glucose, sucrose, fructose, and sorbitol. One day after the first treatment, the second treatment with registered pesticides such as streptomycin (Agrepto), oxytetracycline (Sungbocycline, Sungbo Chemicals Co., Ltd., Republic of Korea), and tri-basic copper sulfate (Saebinna, Syngenta Co., Ltd., Republic of Korea), was carried out. The same procedures were used to assess the impact of secondary treatment on the reduction of the persistence and VBNC state induced by treated streptomycin alone.

Streptomycin absorption was measured as the diameter of the growth inhibition zone of normal E. amylovora cells after they were extracted using E. amylovora persister cells (Chen et al., 2019). Briefly, 1 mL of persister cells induced by streptomycin with 10 mM carbon sources or second treatments with antibiotics were centrifuged at 4,000 rpm for 6 min, washed twice with 10 mM MgCl₂ buffer, resuspended in 100 μL of cell wall digestion buffer [30 mM Tris–HCl (pH 8.0), 1 mM EDTA, 1 mg/mL lysozyme], and incubated at ambient temperature for 2 h. The pellets were frozen by treating them three times with liquid nitrogen. Subsequently, they were denatured for 10 min at 90°C using heat block and centrifuged at 12,000 rpm for 3 min to separate the debris and supernatant. A 10 μL supernatant was used to inoculate a spot in the middle of the pre-inoculated MGY medium with normal E. amylovora cells (100 μL of OD_{600 nm} = 0.1) and incubated at 28°C for 24 h. The diameter of cell growth inhibition zone was measured in two repeats.

An in vitro test was used to confirm that the persister and VBNC cells had recovered after the second antibiotic treatment following streptomycin treatment. Following the second treatment, 1 mL of cells were centrifuged at 4,000 rpm for 6 min and then washed twice in 10 mM MgCl₂ buffer. Thereafter, 50 μL of cells was injected into 5 mL of M9 minimal broth supplemented with 40 mg/mL of asparagine, aspartate, glutamine, tryptophan, thiamine, threonine, histidine, and nicotinic acid and 1.5 mg/mL of succinate and pyruvate. The mixture was then culture at 28°C for 24 h in a shaking incubator. After plating 10 μL of the cultured cells onto MGY medium and incubating them at 28°C for 24 h, the number of bacterial colonies were counted as the number of recovered cells. Three plating cycles were used and three repetitions of recovery test were conducted. King’s B broth served as a positive control and was used in place of the M9 minimal broth supplemented nutrients.

All experiments were independently repeated at least two times. Data were analyzed using t-test and analysis of variance, and the means were compared using Duncan’s least significant range test at p < 0.05. The analysis was performed using the Jamovi project, 2021 (ver. 2.3.21.0).

The degradation rates of different concentration of streptomycin solutions in water at day 7 were not significantly affected, indicating that streptomycin degradation by hydrolysis was not significant (Supplementary Figure 1A). However, the inhibitory zone of 2,500 μg/mL of bactericidal pesticide at day 7 was significantly reduced compared to that at day 0. This result demonstrated that the inhibitory zones of bactericidal pesticide against E. amylovora emerged in a dosage-dependent manner, such that high concentrations (e.g., 2,500 μg/mL) might be degraded to reduce to inhibition size in water under normal light during the 28°C test periods. Streptomycin degradation by UVB, which is a strong factor in photolysis, was not exhibited under the tested conditions in this study (Supplementary Figure 1B). Therefore, we believe that the experimental system has the potential to determine whether E. amylovora can enter dormancy states, including persistence and VBNC, when treated with streptomycin.

Kinetic killing curve assays were used to describe the subpopulation of E. amylovora TS3128 in the presence of 50 and 500 mg/mL concentrations of the laboratory and pesticide streptomycin. In the lag phase, the population declined within 3 h and most of them died (Figure 1A). In contrast, the number of survivors of early and late stationary phase reduced at 4 h and then remained the same until 8 h in both laboratory and pesticide streptomycin treatments (Figures 1B, C). These results suggested that both streptomycin treatments caused TS3128 cells in the stationary phase to enter the culturable persistence state. However, populations of streptomycin-resistant isolate (Psa2) did not decline when treated with streptomycin, and they showed a growth curve that was comparable to King’s B broth without streptomycin (Supplementary Figures 2A, B), indicating that Psa2 was not persistent. The end points of persistence in the kinetic killing curve of King’s B broth and distilled water were determined to be 24 and 48 h, respectively (Figures 1D, E), indicating that the combination of streptomycin treatment and lack of nutrient treatment may induce a longer period of persistence than just streptomycin treatment only.

During the counting of total, live, and culturable cells, culturable cells were observed until days 1 and 2 in King’s B broth and distilled water, respectively (Figures 2A, B). These patterns were linked to those of the kinetic killing curves, indicating that the persistence state occurred early in cells that had contact with streptomycin and would last longer under severe stress condition than in single stress. However, unculturable cells that were considered to be in the VBNC state after the persistence state were analyzed by staining each individual cell according to its physiological characteristics. We were unable to get the five categories that Patel et al. (2021) defined, because in this study, only two fluorescent dyes (the green vitality indicator RSG and red membrane permeability indicator PI) and one mixture fluorescence of them were used. Instead, we identified three classes (Figure 2C) and quadrant (Supplementary Figure 3): class 1, redox-active cells with intact membranes being regarded VBNC cells in this experiment (green, Q4 in FITC-A image); class 2, which are redox-active cells with damaged membranes and considered to be true VBNC cells (yellow, Q2 in FITC-A image); class 3, redox-inactive cells with damaged membranes (red, Q1 in FITC-A image); and no defined class, unstaining cells (white, Q3 in FITC-A image).

Up to 3 and 6 days after being treated with streptomycin in King’s B broth, an average of 34% and 39% of the populations were composed of class 3. After exposure to streptomycin at day 9, class 3 increased by an average of 73%, indicating that many populations changed to damaged membrane cells, which eventually became dead cells (Figure 2D). Unlike in King’s B broth, the populations VBNC cells in class 1 and 2 were the highest, with an average of 71, 70, 74, and 70% for 3, 6, 9, and 15 days, respectively, in streptomycin treated cells in distilled water (Figure 2E). However, class 3 populations did not significantly increase during the long exposure time to streptomycin. Thus, E. amylovora cells were maintained for longer periods as VBNC state in nutrient-deficient condition than in nutrient-rich one.

These findings allowed E. amylovora to experience periods of persistence and VBNC states after exposure to streptomycin, as shown in Figure 3. The persister cells were visible until day 1 before losing their culturability and became VBNC state during days 1–9, and they may have become dead in nutrient-rich condition such as King’s B broth after that time. The persister cells could be seen for up to 2 days in nutrient-deficient conditions such as the distilled water used in this study. After 2 days, they varnished and seemed to be sustained to VBNC cells for an additional 15 days, and they may begin to die after this time. In both conditions, both the persister and VBNC cells were found at those duration times, suggesting that the persistence and VBNC states represent a continuum of dormancy that aids in preventing target activation of streptomycin.

The reduction rates of the persistence and VBNC states induced by streptomycin were evaluated through co-treatment cultivation with carbon sources and second treatment with chemical control agents for fire blight. When rates of stained cells were compared between streptomycin treatment and co-treatments with carbon sources, class 2 rate, which represents the persistence and VBNC state, decreased from 67% (Figure 2E) to 47% in sucrose co-treatment (Figure 4A). Conversely, the rates of class 3, which represents the dead cell caused by membrane damage, increased from 24% to a maximum of 51%, suggesting that sucrose is utilized to activate metabolism-related normal growth. The rates of stained cell types did not change significantly for glucose, fructose, and sorbitol (Figures 4B–D). However, in co-treatments of fructose and sorbitol, the rates of class 1 increased to a maximum of 27–34% at day 7, and populations of culturable cells increased significantly (data not shown), suggesting that these two carbon sources may have initiated re-growth. Thus, we concluded that glucose, fructose, and sorbitol should not be used in formulation of additive composites.

The rates of the three classes in the second treatments of registered pesticides (streptomycin and oxytetracycline) all had the same ratios with streptomycin alone (Figures 5A–C), and the CFUs of culturable cells disappeared at day 3 (data not shown). These results suggest that the two most popular agriculture-antibiotics did not affect the persistence and VBNC states induced by first streptomycin treatment. However, all the types of stained cells quickly changed to class 3 in the second treatment with copper (Figure 5D); thus, we concluded that unculturable cells after 3 days with copper were not in the VBNC state but were all dead cells under in vitro condition. However, two or three treatments with different agricultural antibiotics are recommended as a control technique for fire blight under field conditions. Therefore, according to our results, copper should be used as a treatment once every three treatments.

In the streptomycin absorption experiment, co-treatment of the persister cells with sucrose showed larger inhibitory zone against normal E. amylovora growth than co-treatment of the persister cells with glucose, fructose, and sorbitol (Table 1), and this corresponded to the changed ratios of classes 2 and 3. However, the inhibitory zones of the second treatment of antibiotics were significantly smaller in size than those of the carbon source treatment (Table 1), indicating that antibiotics were not activated by proton motive force.

No change in the ratios of classes 2 and 3 in the second antibiotic treatment was observed, and persister cells of the secondary oxytetracycline treatment did not recover after 2 days (Figure 6A) compared to bactericidal streptomycin treatment alone (Figure 6B) and secondary bactericidal streptomycin treatment (Figure 6C), regardless of the types of amino acid and metabolites. Conversely, the recovered cells from the second streptomycin treatment showed comparable CFUs, indicating that the persister cells induced by single streptomycin treatment were not eliminated by second streptomycin treatment (data now shown). Thus, we concluded that streptomycin persister cells did not survive in the presence of oxytetracycline and oxytetracycline treatment because the second application was highly effective in eliminating culturable streptomycin persister cells.