Screening for Novel Beneficial Environmental Bacteria for an Antagonism-Based Erwinia amylovora Biological Control

Erwinia amylovora, the bacterial species responsible for fire blight, causes major economic losses in pome fruit crops worldwide. Chemical control is not always effective and poses a serious threat to the environment and human health. Social demands for eco-sustainable and safe control methods make it necessary to search for new biocontrol strategies such as those based on antagonists. A bacterial collection from different fire blight-free Mediterranean environments was tested for antagonistic activity against Spanish strains of E. amylovora. Antagonistic assays were carried out in vitro in culture medium and ex vivo in immature loquat and pear fruits. Results revealed that 12% of the 82 bacterial isolates tested were able to inhibit the growth of several strains of the pathogen. Some of the isolates also maintained their antagonistic activity even after chloroform inactivation. Selected isolates were further tested ex vivo, with several of them being able to delay and/or reduce fire blight symptom severity in both loquats and pears and having activity against some E. amylovora strains. The isolates showing the best antagonism also produced different hydrolases linked to biocontrol (protease, lipase, amylase, and/or DNAse) and were able to fix molecular nitrogen. Based on this additional characterization, four biocontrol strain candidates were further selected and identified using MALDI-TOF MS. Three of them were Gram-positive bacteria belonging to Bacillus and Paenarthrobacter genera, and the fourth was a Pseudomonas strain. Results provide promising prospects for an improvement in the biological control strategies against fire blight disease.


Introduction
The phytopathogenic bacterial species Erwinia amylovora is the causative agent of the fire blight of rosaceous plants. The disease affects ornamental species of this family and fruit trees of great economic importance, such as apple, pear, or loquat, as well as some ornamental and wild species such as Cotoneaster, Crataegus, Cydonia, Pyracantha, and Sorbus, among others [1,2], and the disease can be easily spread by rain, wind, insects, birds, and contaminated pruning tools [3][4][5][6]. The first symptoms appear in spring when the bacterium reaches flowers and grows epiphytically on the flower stigmas. Rain and heavy dew wash the pathogen cells to the hypanthium where they enter the plant through the nectarthodes [7,8]. Additional infection routes include other natural openings in young shoots and leaves, such as stomata, hydathodes, trichomes, and/or wounds caused by wind, hail, and insects [7,9]. Once inside the host, the pathogen keeps multiplying and moving through the parenchyma and the xylem vessels, establishing systemic infections [10,11]. Every plant organ is susceptible to infection, although young tissues are more susceptible than lignified, older tissues [4,9]. The main fire blight symptoms include necrosis and secretion of ooze droplets on the affected organ's surface [9]. These ooze droplets contain high concentrations of the pathogen, and their dissemination to susceptible plant tissues

In Vitro Antagonistic Activity: Cross-Streak and Double-Layer Agar Assays
The potential antagonistic activity of the 82 bacterial isolates was initially tested by a modified cross-streak method [40]. Cultured cell suspensions were adjusted to O.D. 600 0.1 ± 0.01 and diluted one-tenth for E. amylovora [10 7 CFU (colony-forming units)/mL] and O.D. 600 0.2 ± 0.02 (10 8 CFU/mL) for potential antagonistic candidates, respectively, in sterile 10 mM Phosphate Buffered Saline (PBS) with pH = 7.2. Briefly, an adjusted suspension of each candidate was individually inoculated, by drawing a line on NA plates and incubated at 28 • C for 48 h. Then, all four pathogenic E. amylovora strains were co-inoculated on the same agar plate on a straight perpendicular line. Each candidate was assessed in duplicate against the four E. amylovora strains. For control treatment, E. amylovora strains were cultured alone, and all plates were incubated at 28 • C for 48 h. Based on these assays, the 10 candidates with better results were selected for further analysis. To determine whether these candidates maintained their antagonistic activity after their biological inactivation, the same assay was repeated including a 15-min chloroform exposure step of the antagonists before streaking the E. amylovora strains. When the ability of the candidates to inhibit E. amylovora growth is due to the production of antimicrobial substances diffused into the culture medium, they maintain their activity after cell inactivation. When the ability of the candidates to inhibit E. amylovora growth is due to competition for space and/or nutrients in the medium, the inhibitory effect is no longer observed after cell inactivation.
The antagonistic activity was also assayed by a modified version of the double-layer agar method [41] to observe E. amylovora growth inhibition halos around the growth area of the tested isolate. In short, 10 µL of a suspension of each candidate adjusted to 10 8 CFU/mL was inoculated onto the center of the plate and then incubated at 28 • C for 48 h. Afterwards, a second layer of top agar was added, which was prepared by mixing 4.5 mL of the agar, melted, and tempered, with 0.5 mL of the suspension of the E. amylovora strain adjusted to 10 8 CFU/mL and incubated at 28 • C for 48 h. The appearance of a halo around the candidate was considered positive for antagonism against E. amylovora.

Ex Vivo Antagonistic Activity: Assays on Detached Fruit
Ex vivo assays were carried out in immature fruits of two sensitive E. amylovora host plant species: loquats (Eriobotrya japonica cv. Tanaka) of 2.3-2.8 cm diameter, and pears (Pyrus communis cv. Williams) of 2.1-2.7 cm diameter. Detached fruits were thoroughly surface-disinfected by immersion in 2% NaClO (v/v) for 5 min, rinsed three times with sterile distilled water, and carefully left to dry on sterile filter paper in the hood. Before inoculation, three equidistant wounds were made on each fruit with a sterile pipette tip, as described elsewhere [42].
Each wound was inoculated with 10 µL of the suspension of each candidate (ca. 10 6 CFU/wound) for 24 h before adding another 10 µL of E. amylovora suspension (ca. 10 5 CFU/wound). Fruits were incubated at 28 • C and in high relative humidity in wet chambers, as described previously [42]. Negative and positive fire blight symptom controls inoculated with PBS and E. amylovora, respectively, as well as controls inoculated only with the antagonists separately, were also included in the assays. Additionally, fruits inoculated with Enterobacter cancerogenus T2-27 were used as positive antagonistic activity control [43]. Each antagonistic strain was tested in four or three replicate fruits, in two independent experiments.
After inoculation, fruits were monitored for fire blight symptom development over time. The "Efficacy" of fire blight symptom control [E (%)] was achieved by the potential candidates, and the "Severity" of the disease [S (%)] was calculated as follows: where "ic" is the number of wounds inoculated with PBS, "it" is the number of symptomatic wounds in the treatment, "T" is the total number of inoculated wounds, "SI it " is the symptom Severity Index of a given inoculated wound, rated according to a visual scale of the symptoms from 0 to 3 as described by [44], and "SI max " is the maximum SI value; in this case, the maximum value is 3. These two parameters, "E" and "S", were evaluated at 5, 7, and 9 days post-inoculation (dpi) in Tanaka loquats and at 3, 5, and 7 dpi in Williams pears, respectively [45]. Following ex vivo assays, re-isolations of E. amylovora from the tested immature fruits were performed on a semi-selective CCT medium, which was PCR-identified by amplification of the 16S rRNA gene, using G1f/G2r primers as described [46].

Exoenzymatic Activities
The ability to hydrolyze proteins was detected as transparent halos around the colonies in Casein Agar medium [NA supplemented with 10% casein (v/v)] according to [47]. Clearing halos around the growth zone are indicative of protease production. DNase activity was assayed in DNase Agar medium (DNA 2 g/L, casein peptone 1.5 g/L, soy peptone 0.5 g/L, NaCl 5 g/L, agar 15 g/L, pH 7.2-7.4) [48]. Plates incubated at 28 • C for 48 h were filled with 1 M HCl to reveal the presence of transparent halos around the growth area due to DNA degradation. Cellulase production was assayed in Cellulose Agar medium (0.5 g/L KH 2 PO 4 , 0.25 g/L MgSO 4 , 2 g/L cellulose, 2 g/L gelatin, 15 g/L agar, pH 7.0) as described by [49]. Plates incubated at 28 • C for 48 h were washed with distilled water and were flooded with 0.1% (w/v) Congo-Red solution in 20% ethanol for 15 min. The liquid was removed, and the plates were decolorized by adding 1 M NaCl solution to reveal the activity. The presence of halos on the discolored medium indicates cellulose hydrolysis. Amylase production was assayed in Starch Agar medium (beef extract 3 g/L, NaCl 2 g/L, dissolvable starch 20 g/L, agar 12 g/L) [50]. To reveal starch degradation, a Lugol solution containing 0.1 M I 2 and 0.2 M KI was used, which only reacts with starch, giving rise to a purple coloration. The amylolytic activity was differentiated by the appearance of transparent halos. Lipase production was assessed in Tween 80 Agar medium (peptone 10 g/L, NaCl 5 g/L, CaCl 2 0.1 g/L, Tween 80 10 g/L, agar 15 g/L) [47]. Degradation of Tween 80 will release free fatty acids, which will react with the Ca +2 ions, generating opaque calcium salts, which precipitate around the lipase-producing colonies.

Identification of Bacterial Strains by MALDI-TOF MS
After the characterization of the bacterial isolates, a selection of the candidates with the best biocontrol potential, supported by the presence of some additional biotechnological activities, was performed for subsequent identification. This was carried out by using Matrix Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry (MALDI-TOF MS) analyses, following the protocol recommended by Bruker Daltonics [52], at the Spanish Type Culture Collection (CECT) of the Universitat de València (Spain). Protein profiles were obtained for each of the bacterial isolates from their ribosomal proteins and were subsequently compared with the protein profiles present in reference databases. Purified cultures grown in NA and incubated at 28 • C for 48 h were prepared for the analyses. The Microflex L20 Mass Spectrometer and the BioTyper 3.1 program were used for data obtention and processing.

Statistical Analysis
In the ex vivo antagonistic assays, a two-way analysis of variance (two-way ANOVA) was used to compare results obtained with the bacterial candidates and the positive controls (E. amylovora CFBP 1430, IVIA 1554, IVIA 1614.2 and IVIA 1892.1) in two independent experiments carried out under the same conditions. The statistical analysis of the data was performed with the GraphPad Prism 9 software. In all analyses, a p value < 0.05 was considered statistically significant. All results were represented as the mean ± the standard deviation (SD).

Results and Discussion
In many parts of the world, the control methods in crop protection against E. amylovora are frequently based on the preventive but not curative use of antibiotics such as streptomycin and/or copper-based compounds. However, the associated emerging antibiotic resistance has led the EU to ban their use in plant protection products [30,53], and the phytotoxicity issues related to the application of copper have resulted in major restrictions on its use [24,54]. The importance of finding an environmentally friendly and effective control mechanism against fire blight due to the lack of efficient phytosanitary control measures to stop its dissemination makes E. amylovora a pathogen of great concern worldwide [55]. In recent years, biological control with microbial antagonists has been considered a powerful and eco-friendly alternative [56]. In this panorama, the aim of this work was to evaluate the potential antagonistic activity of a collection of environmental bacterial isolates against E. amylovora.

E. amylovora-Free Environments Can Be a Source of E. amylovora Antagonists
Based on the in vitro cross-streak assay, 10 of the 82 bacterial isolates tested were selected as good candidates for further characterization. These were candidates UV-17, UV-18, UV-19, UV-20, and UV-41, which inhibited the growth of all the four tested E. amylovora control strains, and candidates UV-9, UV-12, UV-30, UV-59, and UV-60, which inhibited the growth of between one and three of the tested E. amylovora strains ( Table 2).
Active Candidates UV-12 and UV-30 were not able to maintain their antagonistic activity against the pathogen after their inactivation with chloroform. This indicates either potential contact-dependent antagonistic activity, quick inactivation of released antimicrobial compounds under the assayed conditions, and/or active modification of the medium pH while feeding on the media, although more tests are required to confirm these possibilities [57]. Most of the antagonist candidates listed in Table 2 showed E. amylovora inhibitory effects with and without inactivation, suggesting the production of stable antimicrobial substances that diffused through the medium and remained active during the assayed period.
Based on the results, candidates UV-12, UV-17, UV-18, UV-19, UV-20, UV-30, and UV-41 were selected due to their stronger antagonistic activity [higher activity values (++/+++)] against three or more E. amylovora control strains, despite some of them (UV-12 and UV-30) completely lost their function after chloroform inactivation [24]. A strong antagonistic activity against this pathogen has already been described by the use of plant-associated antagonists that compete for the scarce nutrients in the medium [43]. Representative pictures of the cross-streaking method results are shown in Figure 1.
In vitro antagonistic activity of the bacterial isolates was also tested by the double-layer agar method as described by [41]. However, inhibition halos of E. amylovora could not be quantified, then no comparative readings could be made using this method.

Ex Vivo Antagonistic Activity on Detached Fruit
The biocontrol potential of the ten candidates selected in vitro was also assessed ex vivo on immature fruits of two host species susceptible to E. amylovora: Tanaka loquats and Williams pears. In this case, the candidate lost its potential biocontrol activity against some E. amylovora strains afte chloroform inactivation, suggesting that the growth inhibitory activity may be due more to compe tition for space and/or nutrients in the medium than to the production of antimicrobial substances In vitro antagonistic activity of the bacterial isolates was also tested by the double layer agar method as described by [41]. However, inhibition halos of E. amylovora coul not be quantified, then no comparative readings could be made using this method.

Ex Vivo Antagonistic Activity on Detached Fruit
The biocontrol potential of the ten candidates selected in vitro was also assessed e vivo on immature fruits of two host species susceptible to E. amylovora: Tanaka loquats an Williams pears.

Antagonistic Activity of the Candidate Strains against Fire Blight in Loquats
An initial test was conducted as a set-up of the experiment. Candidates UV-9, UV-59 and UV-60, which had already been discarded for their low in vitro activity (Table 2), wer also discarded for their low ex vivo activity against E. amylovora. Two independent exper iments were conducted on immature loquats with the new selection of candidates (UV 12, UV-17, UV-18, UV-19, UV-20, UV-30, UV-41, and the infection inhibition control E cancerogenus IVIA T2-27) and 1/10 dilutions of the E. amylovora control strains. Results ar summarized in Figure 2. Regarding Tanaka loquats inoculated with the positive E. amylo vora inhibition control IVIA T2-27, they showed 76% control of symptom developmen the highest in all cases. Further, in the few inoculated wounds showing symptoms, th severity was also the lowest throughout the experimental period. The results confirme the antagonistic activity of strain IVIA T2-27 from loquat microbiota against the E. amylo vora strains [43]. . In this case, the candidate lost its potential biocontrol activity against some E. amylovora strains after chloroform inactivation, suggesting that the growth inhibitory activity may be due more to competition for space and/or nutrients in the medium than to the production of antimicrobial substances.

Antagonistic Activity of the Candidate Strains against Fire Blight in Loquats
An initial test was conducted as a set-up of the experiment. Candidates UV-9, UV-59, and UV-60, which had already been discarded for their low in vitro activity (Table 2), were also discarded for their low ex vivo activity against E. amylovora. Two independent experiments were conducted on immature loquats with the new selection of candidates (UV-12, UV-17, UV-18, UV-19, UV-20, UV-30, UV-41, and the infection inhibition control E. cancerogenus IVIA T2-27) and 1/10 dilutions of the E. amylovora control strains. Results are summarized in Figure 2. Regarding Tanaka loquats inoculated with the positive E. amylovora inhibition control IVIA T2-27, they showed 76% control of symptom development, the highest in all cases. Further, in the few inoculated wounds showing symptoms, the severity was also the lowest throughout the experimental period. The results confirmed the antagonistic activity of strain IVIA T2-27 from loquat microbiota against the E. amylovora strains [43].
With respect to fruits inoculated with the selection of candidates, significant antagonistic activity was confirmed at 5 dpi. However, 100% efficacy was not achieved in any case, and infection control decreased along with incubation time (Figure 2a-d). Some candidates were able to notably reduce the severity of the symptoms compared to the positive infection controls. In general, fruits inoculated with candidates UV-12, UV-17, and UV-20 achieved greater efficacy against all E. amylovora strains, with significant differences from the positive control at 5 dpi in all cases: candidate strain UV-12 showed a control efficacy ranging from 70-80%, depending on the assayed E. amylovora strain. Candidates UV-17 and UV-20 showed control efficacies of 65-85% and 50-75%, respectively. These strains also showed the greatest capacity to reduce symptom development in the tested loquats throughout the experiment. The efficacy of candidate UV-30 in controlling symptom development and severity was also significantly better than untreated controls when assayed against E. amylovora IVIA 1554 and IVIA 1892.1. The lowest disease severity was observed in fruits inoculated with candidates UV-12, UV-17, and UV-30 regardless of the tested E. amylovora strain (Figure 2e-h).   In the E. amylovora-inoculated fruits (positive fire blight controls), the disease severity at 5 dpi varied according to the pathogenic strain tested: 61% for CFBP 1430 (Figure 2e), 53% for IVIA 1614.2 ( Figure 2g) and 36% for IVIA 1554 (Figure 2f) and IVIA 1892.1 (Figure 2h). Symptom development was progressive, and at 9 dpi severity was 100% for CFBP 1430, 86% for IVIA 1554, 89% for IVIA 1614.2, and 69% for IVIA 1892.1. No fire blight symptoms were observed throughout the experiment in the fruits inoculated with PBS ( Figure 2) or with the antagonists alone.
Representative pictures of infection and symptoms on the immature Tanaka loquats from these ex vivo antagonistic assays are also shown in Figure 2i,j.

Antagonistic Activity of the Candidate Strains against Fire Blight in Pears
Due to the rapid onset of fire blight symptoms observed in immature pear fruits, the incubation time had to be reduced to 7 days, taking readings at 3, 5, and 7 dpi. Results are shown in Figure 3. Fruits inoculated with E. amylovora 24 h after their inoculation separately with the selection of candidates showed significant control of symptom development at 3 dpi (Figure 3a-d). Fruits inoculated with candidates UV-17, UV-20, and UV-30 showed the best efficacy against E. amylovora strains (Figure 3a-d). With a value of up to 89% in fruits inoculated with E. amylovora IVIA 1614.2 at 3 dpi and 67% at 5 dpi, UV-30 showed the best antagonistic activity among the evaluated candidates (Figure 3c). To a lesser extent, some efficacy was achieved with candidate UV-19, mainly against IVIA 1614.2 and IVIA 1892.1, controlling symptom development in 50% of the inoculated fruits at 3 dpi (Figure 3c,d, respectively).
Despite the relatively good disease control efficacy provided by some of the candidate strains at 3 dpi, a drastic drop in the efficacy took place at 5 dpi (Figure 3a-d). By 7 dpi, fruits inoculated with candidate UV-30 against E. amylovora IVIA 1892.1 were the only ones with some efficacy (6%) (Figure 3d).
Regarding the disease control efficacy on fruits inoculated with the positive antagonistic activity control strain E. cancerogenus IVIA T2-27, it was 89-100% at 3 dpi, the highest in all cases (Figure 3a-d), and the severity was also the lowest during all incubation time (Figure 3e-h). These results confirmed the antagonistic activity of strain E. cancerogenus IVIA T2-27 from loquat microbiota against the assayed E. amylovora strains, according to previous studies [40].
Representative pictures of infection and symptoms on the immature Williams pears from these ex vivo antagonistic assays are also shown in Figure 3i,j. (Figure 3f-h) and candidates UV-18 and UV-19 against E. amylovora strain CFBP 1430 ( ure 3e), with significant differences with the positive control in all cases at 3 dpi. The severity observed in fruits inoculated with candidate UV-17 against IVIA 1614.2 and I 1892.1 was also remarkable (Figure 3g,h, respectively).  Altogether, results revealed candidate UV-12 as the most effective in Tanaka loquats at 9 dpi (above 40%), followed by UV-17 and UV-30 (20 to 40%) and UV-20 (15 to 20%) (Figure 4a). On the other hand, the most effective candidates in Williams pears at 3 dpi were UV-20 and UV-30 (above 50%), followed by UV-17 (40 to 50%) and candidates UV-18 and UV-19 to a lesser extent (20 to 30%) (Figure 4b). Regarding the disease severity, UV-30 was the candidate that most reduced the appearance of symptoms in Tanaka loquats at 9 dpi (below 35%), followed by candidates UV-12, UV-17, and UV-20 (35 to 50%) (Figure 4c). Candidates UV-30, UV-17, and UV-20 were those that better controlled the appearance of symptoms in Williams pears (below 20%), followed by UV-18 and UV-19 (20 to 30%) (Figure 4d).
Representative pictures of infection and symptoms on the immature Willia from these ex vivo antagonistic assays are also shown in Figure 3i,j.
Altogether, results revealed candidate UV-12 as the most effective in Tanak at 9 dpi (above 40%), followed by UV-17 and UV-30 (20 to 40%) and UV-20 (1 (Figure 4a). On the other hand, the most effective candidates in Williams pear were UV-20 and UV-30 (above 50%), followed by UV-17 (40 to 50%) and candid 18 and UV-19 to a lesser extent (20 to 30%) (Figure 4b). Regarding the disease UV-30 was the candidate that most reduced the appearance of symptoms in T quats at 9 dpi (below 35%), followed by candidates UV-12, UV-17, and UV-20 (3 (Figure 4c). Candidates UV-30, UV-17, and UV-20 were those that better contr appearance of symptoms in Williams pears (below 20%), followed by UV-18 an (20 to 30%) (Figure 4d).  This work highlights the influence the environment may have on antagonism. Potential biocontrol agents may not behave the same in a controlled space as shown in in vitro assays, as in ex vivo assays, where many variables (available nutrients, pH, humidity, and temperature conditions) and even the microbiota of the plant material itself can have an influence, making it difficult to standardize. In addition, the differences obtained among potential biocontrol candidates and E. amylovora strains on immature loquats (cv. Tanaka) and pears (cv. Williams) point out the importance of performing ex vivo assays on plant material from different species.
The nitrogen-fixing activity was the growth-promoting activity tested in this work. As shown in Table 3, candidates UV-12, UV-17, UV-20, and UV-30 were able to fix nitrogen at 48 h. Candidates UV-18 and UV-19 showed nitrogen-fixing activity at 96 h. Candidates UV-9, UV-41, UV-59, and UV-60 did not show nitrogen-fixing capacity throughout the whole assay.
A second selection of candidates was made based on the results: candidates UV-12, UV-17, UV-20, and UV-30 displayed the best efficacy against E. amylovora and the lowest disease severity on immature fruits in general. They also showed earlier nitrogen-fixing activity than other candidates and were negative for cellulase activity, which could interfere with planta assays due to its ability to degrade plant material, which has been described in many plant pathogens as a virulence factor [58].

Identification of Bacterial Strains by MALDI-TOF MS
The selected candidates with the highest biocontrol potential, based on the in vitro and ex vivo assays, also supported by the presence of some additional biotechnological activities, were identified by the MALDI-TOF MS methodology (Table 4). Candidates UV-17 and UV-30 were identified as Paenarthrobacter aurescens and Pseudomonas moraviensis, respectively, and candidates UV-12 and UV-20 as species within the genera Bacillus and Paenarthrobacter, potentially B. simplex and P. nicotinovorans, respectively. All three genera correspond to bacteria described in different environmental samples as plant growthpromoting rhizobacteria [57,59]. Results support the existing body of literature that places Bacillus spp. As one of the most studied species for biocontrol. Not only a multitude of direct antagonistic mechanisms against plant pathogens, including hydrolytic enzyme and antibiotic production or competition for nutrients and space but also excellent plant growth-promoting and disease-reducing properties of plants caused by both phytopathogenic fungi and bacteria have been reported [60,61]. B. simplex has previously been described with a wide range of activities related to biocontrol potential, supported by the exoenzymatic production and nitrogen-fixing abilities observed in candidate UV-12 [62,63].
Some species of the genus Paenarthrobacter have been reported to have remarkable metabolic capabilities to degrade xenobiotics used in industrial applications and bioremediation [64]. Candidate UV-20, identified as P. nicotinovorans, has been not only described as having nitrogen-fixing capacity [65], but it is also capable to degrade nicotinic acid, an essential growth factor for E. amylovora [66]. This ability could be decisive to confirm its antagonistic capacity against this phytopathogen. However, to date, only the biocontrol potential of Paenarthrobacter ureafaciens has been previously evaluated and reported [59].
Pseudomonas species stand out for having great metabolic versatility and a broad spectrum of activities related to pathogen biocontrol, such as antibiotic and hydrolase production, and plant growth promotion such as nitrogen and other nutrient fixation [67]. Even though P. moraviensis has not been described as a biocontrol agent to date, its phosphatesolubilizing ability to promote plant growth and the results obtained in this work may make isolate UV-30 a suitable candidate for antagonism against E. amylovora [68].
The efficacy of some antagonistic bacteria for the control of fire blight in detached loquat and/or pear fruit has been demonstrated in recent studies with plant-associated bacterial species [40] and, more recently, with a strain of Priestia megaterium isolated from a pear orchard [56]. However, neither Paenarthrobacter aurescens, P. nicotinovorans nor Pseudomonas moraviensis from fire blight-free Mediterranean environments have been described so far as potential biocontrol agents of E. amylovora. Notwithstanding, in order to evaluate the potential of the selected candidates as biocontrol agents against fire blight, further in vivo studies would be needed to assess their potential to be used in biocontrol under field conditions.

Conclusions
The present study concludes that the selected bacterial candidates have demonstrated in vitro antagonistic ability against E. amylovora through nutrient competition and the production of antimicrobial substances in some cases. Results of antagonism assays on immature fruits of two host species susceptible to E. amylovora revealed the ability of some of them to reduce and delay the onset of fire blight symptoms, with several of them showing hydrolytic and plant growth-promoting activities related to biocontrol potential. These antagonistic bacteria could contribute to providing or improving new, safer, and more environmentally friendly biological control strategies against fire blight disease.