agronomy The Potential of Novel Bacterial Isolates from Natural Soil for the Control of Brown Rot Disease ( Monilinia fructigena ) on Apple Fruits

: Monilinia fructigena is one of the most important fungal pathogens causing brown rot on apple and is heavily a ﬀ ecting fruit production. The main objective of this study was to screen for potential bacterial isolates with higher antagonistic activity against M. fructigena . Our study focused on the identiﬁcation of potential bacterial isolates capable of reducing both the mycelial growth of M. fructigena and the disease severity using in vitro and in planta trials, respectively. To achieve this goal, thirteen bacteria, isolated from natural soil, were evaluated for their abilities to produce lytic enzymes (amylase, cellulase and protease), hydrocyanic acid (HCN) and lipopeptides (bacillomycin, fengycin, iturin and surfactin). Further, results from the dual culture method, volatile and bacterial free-cell ﬁltrate bioassays indicated that tested isolates showed a fungicidal activity against the mycelial growth of M. fructigena . Thus, out of the 13 isolates tested, 12 exhibited signiﬁcant mycelial inhibition (more than 70%) against M. fructigena, while remaining the last isolate displayed only a partial inhibition (up to 43%). Further, 12 of the bacteria isolates displayed an amylase production, 10 showed cellulase production, 11 revealed protease production, while only 2 displayed HCN production. In addition, most bacterial isolates were found to have genes encoding for di ﬀ erent lipopeptides: bacillomycin (10), fengycin (3), iturin (11) and surfactin (1). Interestingly, two bacterial isolates, Bacillus amyloliquefaciens B10W10 and Pseudomonas sp. B11W11 were found to be the most e ﬀ ective and displayed the lowest disease severity in planta trial. These two bacteria reduced the brown rot incidence compared to the synthetic fungicide in a semi-commercial large-scale trial. Therefore, our ﬁndings suggest that these two later bacterial isolates provide apple protection against M. fructigena via direct and indirect mechanisms. These isolates may be used, therefore, as potential biological control agents (BCAs) in preventive treatment to control brown rot disease on apple fruits.

Agronomy 2020, 10, 1814 3 of 19 To improve biocontrol strategies for pathogens that affect apple fruits in Morocco, there was an increase of interest among researchers to develop bacterial-based fungicides. The intended purpose associated with these types of pesticides is to ensure the control of fungal diseases and the enhancement of the BCAs efficacy and reliability. Therefore, the main objectives of this study were (i) screening, characterization and identification of novel potential bacterial isolates; (ii) assessing their biocontrol efficacy against M. fructigena, both in vitro and in vivo; (iii) studying their modes of action; and (iv) finally proposing the most effective bacterial isolates for brown rot control.

Fungal Pathogen Preparation
Monilinia fructigena (strain VPBG) used in this study was isolated from cherries in Serbia and identified as described previously [21]. The fungal pathogen was subcultured on potato dextrose agar (PDA) medium and incubated 7 days in darkness at 25 • C using an IN30 cultivator instrument (Memmert GmbH Co., Koln, Germany).
The conidial suspension was prepared by flooding a 10-14-day fungal colony with 5 mL of sterile distilled water (SDW) containing Tween-20 (0.05%). The conidial suspension was recovered by scraping the surface of the medium with the tip of a sterile Pasteur pipette. The resulting suspension was filtered through autoclaved tissue layers to remove debris, mycelium and medium. The concentration of conidial suspension (1 × 10 6 and 1 × 10 3 spores/mL) was adjusted determined using a Malassez hematocytometer (Roche, Meylan, France) by adding SDW.

Bacterial Isolates: Preparation and Molecular Identification
Thirteen bacterial isolates from the culture collection of "Ecole Nationale d'Agriculture of Meknes" (ENA-Meknes), Morocco, and previously isolated from the natural, uncultivated calcimagnesic soil of ENA-Meknes were screened for their potential antagonistic activity against M. fructigena. Before experiments, each bacterial isolate was maintained on Luria-Bertani (LB) medium at 25 • C for 48 h. The bacterial suspension was prepared from a 24 h-old culture grown on PDA medium at 25 • C and adjusted to 1 × 10 8 CFU/mL concentration using a spectrophotometer at 620 nm wavelength.
Genomic DNA of each bacterial isolate was extracted from fresh bacterial suspensions according to the DNA extraction protocol developed by LIop et al. [22]. Bacteria were identified based on their partial 16S ribosomal DNA gene using the primer pair Fd1/RP2 (Fd1:5 -AGAGTTTGATCCTGGCTCAG-3 and RP2:5 -ACGGCTACCTTGTTACGACTT-3 ). The polymerase chain reaction (PCR) was carried out in a total volume of 25 µL using EnzimaGoTaq DNA polymerase (Bioline, London, UK) according to the manufacturer instructions. Each PCR test consisted of 5 µL (Buffer 5x), 0.25 µL Taq, 1 µL forward Primer (50 µM), 1 µL reverse Primer (50 µM), 15.25 µL of H 2 O and 2.5 µL of DNA template of each bacterial isolate. The negative control contained only the reaction mixture and SDW. PCR amplifications were carried out in an Eppendorf thermal cycler following conditions: initial denaturation at 96 • C for 4 min followed by 35 cycles of denaturation at 96 • C for 10 s, primers annealing at 52 • C for 40 s and extension at 72 • C for 2 min; with a final extension step at 72 • C for 4 min. PCR products were visualized by agarose gel electrophoresis (1% agarose) and stained with ethidium bromide. Sequencing of PCR-amplified DNA was performed. The obtained sequences were edited and aligned using DNAMAN software (version 6.0, Lynnon Biosoft, Quebec Canada) and compared for homology in the NCBI-BLAST database (National Center for Biotechnology Information http://blast.ncbi.nlm.nih.gov/Blast.cgi). Sequences resulting from this analysis were deposited in GenBank under accession numbers, as mentioned in Table 1.

Biochemical Traits of Bacterial Isolates
The ability of the thirteen bacterial isolates to produce the amylase activity was carried out on Petri dishes containing the nutrient agar (NA) medium (10 g beef extract, 10 g peptone, 10 g NaCl, 20 g glucose, 15 g agar and 1000 mL of SDW) supplemented with 1% of soluble starch [23]. Briefly, each 48 h-old bacterial culture was spotted (10 µL) onto the center of the Petri dish and incubated for 72 h at 28 • C. Three ml of the iodine solution was added to each Petri plate to visualize the starch hydrolysis. The appearance of a clear zone (halo) around the colony demonstrates a positive amylase activity, whereas a negative activity was characterized by the occurrence of brown color around the colony [24]. For the proteolytic activity, the ability of the thirteen bacterial isolates to produce protease was evaluated on skimmed milk agar medium as previously described by Kumar et al. [25]. An aliquot of 5 µL of each 24 h-old bacterial culture was deposited in the center of each Petri plate. Petri dishes were then incubated for 72 h at 28 • C. The formation of a clear zone around colonies indicated a positive proteolytic activity. In addition, the activity of the cellulase of all bacterial isolates was determined using carboxymethyl cellulose (CMC) solid medium [26]. Briefly, an aliquot (10 µL) of 24 h-old bacterial culture was placed in the center of the Petri plate. The cellulase activity was determined after 8 days of incubation period at 28 • C by adding 0.1% Congo red dye to each Petri dish for 15 min, rinsed three times with 1 M NaCl solution and examined for formation of a clear zone (halo) around colonies. In the last, the ability of the isolates to produce hydrocyanic acid (HCN) was evaluated according to the protocol developed by Bakker and Schipperes [27] with slight modifications. For each bacterial isolate, a 24 h-old culture was prepared and adjusted to a concentration of 1 × 10 8 CFU/mL using a spectrophotometer. An aliquot (100 µL) of each bacterial suspension was inoculated into Petri dish containing yeast extract peptone dextrose (YEPD) medium (10 g glycose, 5 g peptone, 5 g yeast extract and 15 g agar) amended with glycine (4.4 g/L). A sterile Whatman filter paper no.1 soaked in a picrate solution (2.5% picric acid and 12.5% anhydrous sodium carbonate (Na 2 CO 3 )) was placed at the top of each Petri dish. The plates were sealed with Parafilm, inverted, then incubated at 28 • C for 4 days. A control Petri dish only inoculated with SDW served as a negative control. Response was assessed by visual observation of the color of Whatman paper. The change of the color from yellow to orange or brown indicates the production of volatile HCN.

In Vitro Dual Culture Bioassay
Dual culture plate bioassay was used to screen the best antagonist bacterial isolates against M. fructigena. A twenty-four-hour-old bacterial culture of each isolate was spotted at 4 equidistant streaks (3 to 4 cm length) along the perimeter of the Petri dish containing the PDA medium. A mycelial plug of 7 mm in diameter was cut out from the edge of the 7-day-old actively growing colony, deposited in the center of each Petri dish within bacterium bands and incubated at 28 • C [31]. Petri dishes with only mycelial plugs served as a negative control. The antagonistic activity was determined by calculating the inhibition rate (IR) of mycelial growth as follows: where D C is the diameter of the colony in control treatments while D T corresponds to the diameter of the colony in confrontation treatments. The experiment was repeated twice over time, with 4 Petri plates for each bacterial isolate.

Volatile Products (VOCs) Bioassay
The antifungal activities of bacterial isolates through the production of VOCs were evaluated by in vitro bioassays. To do this, each bacterium was first subcultured into three parallel streaks on Luria-Bertani medium and incubated at 28 • C for 24 h. Afterward, the lid of the Petri plate was replaced by the bottom of another Petri plate containing a 7 mm-mycelial plug from the edge of the 10-day-old actively growing fungal colony on PDA. Both bottoms plates were sealed together with Parafilm [32]. Control plates were prepared in the same way, without bacterial isolates. The plates were incubated at 28 • C, and inhibitions rates were recorded as described above after 7 days of incubation periods. This experiment was repeated twice over time, with 4 replicates for each treatment.

In Vitro Bacterial Cell-Free Filtrates Effect
Each bacterial isolate was first subcultured on nutrient broth (NB) medium at 28 • C for 48 h. An aliquot (100 µL) of each bacterial suspension (1 × 10 8 CFU/mL) was inoculated in an NB culture medium containing 10 g beef extract, 10 g peptone, 10 g NaCl, 20 g glucose and 1000 mL of SDW and then incubated in a rotary shaker (130 rpm) at 30 • C for 3 days. The bacterial cell cultures were centrifuged for 25 min (5.500 rpm), and the supernatant of each bacterium was recovered and filtrated by using a 0.22 µm Millipore filter [33]. The obtained cell-free filtrate was used for in vitro dual culture tests at a concentration of 10% [33]. In addition, a negative control was prepared by adding liquid NA medium to the PDA medium instead of cell-free culture filtrate. Petri plates were then inoculated by 7 mm-mycelial plugs from the edge of the 10-day-old actively growing colony and incubated at 28 • C. The diameter of mycelial growth was determined at 5 and 10 days post-inoculation and used to calculate the inhibition rates as described above. This experiment was repeated twice over time, with four replicate for each bacterium treatment.
To determine the impact of bacterial cell-free filtrates on spore germination, each bacteria-free filtrate (100%) was mixed with conidial suspension of Monilinia (1 × 10 6 ) spores/mL at a ratio of 1:1 (v:v). In addition, a bacteria-free NA culture solution mixed with spore suspension at the same volume was served as controls. All mixtures were incubated at 25 • C in Eppendorf tubes. Germination of the spores was observed under the microscope after 24 h of incubation. A spore is considered germinated if the length of the germ tube is greater than its smaller diameter. The inhibition rate of spore germination (Ig) is determined according to the following formula: with N C : number of germinated spores in culture medium without bacterial cell-free filtrate and N T : number of germinated spores in bacteria cell-free filtrate (100%).

In Vivo Antagonism Experiment
Apple fruits (Starkrimson) were used for in vivo experiments. Apple fruits with uniform size, a similar maturity stage and without blemishes, injuries or visible rot were stored at 4 • C prior to until use for experiments. Prior to each experiment, apple fruits were surface disinfected by soaking in the sodium hypochlorite solution (1%) for 5 min, washing twice with SDW and drying for 1 h at room temperature under the laminar flow hood.
To monitor the biocontrol efficacy of bacterial isolates on apple fruits, two equidistant wound sites (4 mm in diameter and 3 mm in depth) were performed on the equatorial zone of the fruit using a sterile stainless-steel rod. Each apple wound was treated with 50 µL of each bacteria suspension (1 × 10 8 CFU/mL) and inoculated 4 h later with 50 µL of fungal suspension of M. fructigena (1 × 10 3 and 1 × 10 6 spores/mL). Wounds that received only SDW were served as controls. Each inoculated apple was deposited in a disinfected cylindrical plastic box (1 L of capacity) where a Whatman paper soaked with SDW was placed to maintain a high humidity level for fungal growth and to prevent moisture loss from the fruit. All boxes were incubated at 24 • C in a humid growth chamber with a 12 h photoperiod. The experiment was repeated twice over time. For each treatment (4 fruits, 8 wounds), the lesion diameters were measured after 5 and 10 days post-inoculation and the disease severity (DS) of pathogenic fungus on apple fruits was calculated according to the following formula: DS (%) = (average lesion diameter (mm) of treated wounds/average lesion diameter (mm) of the wounds in the control treatment) × 100 For the semi-practical trials of selected bacteria, out of the 13 tested bacterial isolates, two antagonistic bacteria displaying low disease severity in vivo and the highest inhibition rates of mycelial growth in vitro were selected for a semi-commercial large-scale trial. Apple fruits were wounded (3 Agronomy 2020, 10, 1814 7 of 19 mm in diameter and 4 mm-deep) at four equidistant points. They were subsequently dipped in each bacterial suspension (1 × 10 8 CFU/mL) for 2 min. Apple fruits dipped in SDW served as a control. Apple fruits were placed in plastic bags (10 apple fruits/bag) with three replicates and incubated in the growth chamber at 24 • C. After 24 h of the incubation period, apple fruits were inoculated by spraying with M. fructigena (1 × 10 3 spores/mL) and incubated in a growth chamber at 24 • C. The number of infected fruits was monitored every 5 days until apples in control treatments were totally infected. The experiment was repeated twice over time. The disease incidence (%) was calculated for each treatment according to the following formula: Incidence (%) = number of infected fruits/total number of fruits × 100

Statistical Analysis
All in vitro and in vivo experiments were carried out twice over time. Analysis of variance (ANOVA) was performed using SPSS statistical software (version 20). When the effect was revealed to be significant, the least significant difference (LSD) test was performed for means separation at p = 0.05.

Bacterial Identification
The bacterial 16S rRNA sequences obtained were deposited in the GenBank database of NCBI with the accession numbers listed in Table 1. Five isolates were identified as B. subtilis. In addition, four bacterial isolates were identified as B. amyloliquefaciens. One isolate was classified as B. siamensis, while the last three isolates were found to belong to Pseudomonas sp.

Biochemical Traits of Bacterial Isolates
The ability of all bacterial isolates to hydrolyze starch showed that 92.31% of isolates exhibited an amylase activity ( Figure 1 and Table 3). However, a difference in amylase production between the tested bacterial isolates was observed. Therefore, the bacterial isolate B. amyloliquefaciens B4W4 showed the most important activity (1.4), followed by B. amyloliquefaciens B7W7, while no amylolytic activity was exhibited by Pseudomonas sp. strain FD1-9F. Table 3. Capacity of bacterial antagonists to produce lytic enzymes involved in the biocontrol mechanisms such as amylase, protease, cellulase and hydrocyanic acid (HCN) production.

Bacterial Strains
Amylase Protease Cellulase HCN  Results of protease activity showed that out of 13 tested bacteria, 11 (84.62%) were able to exhibit the protease activity ( Figure 1 and Table 3). Indeed, isolate B. amyloliquefaciens B4W4 had the highest activity (2.5), whereas both isolates Pseudomonas sp. strain FD1-9F and Pseudomonas sp. strain B11W11 do not show any protease activity.
The evaluation of the ability of isolates to hydrolyze the cellulose revealed that 76.92% of strains have a positive cellulosic activity ( Figure 1 and Table 3). However, there was a difference between bacterial isolates in synthesizing cellulase molecules, with then isolates having a positive activity varying between 1 and 1.38. Moreover, only Bacillus subtilis B5W5 had the highest index (1.38), followed by isolate RP2-13F with an index of 1.29. On the other hand, no cellulase activity was recorded for Pseudomonas sp. B6W6 and B. amyloliquefaciens B7W7.
The analysis of HCN production showed that only two bacterial isolates out of 13 (15.38%) were able to produce HCN ( Figure 1 and Table 3). However, these two bacterial isolates have only a moderate HCN production.

Detection of Lipopeptides by PCR
The PCR results for the detection of genes involved in the production of lipopeptides in bacterial antagonists are listed in Table 4. Table 4. Detection of genes involved in the production of lipopeptides in bacterial antagonists. Results of protease activity showed that out of 13 tested bacteria, 11 (84.62%) were able to exhibit the protease activity ( Figure 1 and Table 3). Indeed, isolate B. amyloliquefaciens B4W4 had the highest activity (2.5), whereas both isolates Pseudomonas sp. strain FD1-9F and Pseudomonas sp. strain B11W11 do not show any protease activity.

Strain Bacillomycin Fengycin Iturin Surfactin
The evaluation of the ability of isolates to hydrolyze the cellulose revealed that 76.92% of strains have a positive cellulosic activity ( Figure 1 and Table 3). However, there was a difference between bacterial isolates in synthesizing cellulase molecules, with then isolates having a positive activity varying between 1 and 1.38. Moreover, only Bacillus subtilis B5W5 had the highest index (1.38), followed by isolate RP2-13F with an index of 1.29. On the other hand, no cellulase activity was recorded for Pseudomonas sp. B6W6 and B. amyloliquefaciens B7W7.
The analysis of HCN production showed that only two bacterial isolates out of 13 (15.38%) were able to produce HCN ( Figure 1 and Table 3). However, these two bacterial isolates have only a moderate HCN production.

Detection of Lipopeptides by PCR
The PCR results for the detection of genes involved in the production of lipopeptides in bacterial antagonists are listed in Table 4. Table 4. Detection of genes involved in the production of lipopeptides in bacterial antagonists.

Strain Bacillomycin Fengycin Iturin Surfactin
Bacillus amyloliquefaciens B1W1 Table 4. Cont.  (Table 4). For the fengycin production, two primers FEND1-F/FEND1-R, were used for the detection of the fenD gene in the genomic DNA of bacterial isolates. The results revealed bands of 964 bp in 3 bacterial isolates out of 13 (Table 4). Bacterial isolates for which the fenD gene was detected were Pseudomonas sp. FD1-9F, Pseudomonas sp. B11W11 and B. subtilis RP2-12F. However, for iturin secretion, the detection of the iturin operon, responsible for the biosynthesis of iturin, two primers were used ITUP1-F and ITUP1-R, which showed 2 kb fragments containing parts of the iturin A and iturin B genes and intergenic sequences between these two genes. Among the 13 tested bacteria, 2 bacterial isolates Pseudomonas sp. B6W6 and Pseudomonas sp. B11W11 lack this gene (Table 4). Lastly, the surfactin production was evaluated using two primers P17/ P18 to detect the presence of the sfp gene encoding for the 4 -phosphopantetheine transferase, which is required for the activation of surfactin synthetase. Results indicated that only isolate Pseudomonas sp. B6W6 had the sfp gene (Table 4).

In Vitro Dual Confrontation Bioassay
A dual culture method was used to investigate the potential antagonistic effect of the 13 bacterial isolates against the pathogenic fungus M. fructigena (Table 1). Results revealed that out of the 13 tested isolates, 12 showed a significant inhibition (more than 70%) against M. fructigena, while the remaining isolate displayed only a partial inhibition (up to 43%) after 10 days of incubation. Further, the statistical analysis highlighted a highly significant difference between isolates. Overall, the mycelial growth in the presence of bacterial isolates was significantly reduced than that of controls. It was noticed, however, that mycelial growth inhibition rates increased with the increasing incubation time from 5 to 10 days (Table 1).

Effect of Bacteria-Derived VOCs on Mycelial Growth, under In Vitro Conditions
The in vitro tests of VOCs derived from the 13 tested bacterial isolates revealed that 10 isolates have antagonistic activity against M. fructigena by producing volatile antifungal compounds, but with a low IR (35%). In addition, three isolates have an activity of less than 20% after 5 days of incubation; further, their activities disappeared after 10 days post-incubation ( Figure 2). Overall, statistical analyses displayed a highly significant difference between isolates.

Effect of Bacterial Cell-Free Filtrates on Mycelial Growth and Spore Germination
Inhibition rates of mycelial growth of M. fructigena obtained with bacterial cell-free filtrates were highly significantly different between all the 13 isolates ( Figure 3). e cell-free filtrates of both bacterial isolates B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11 have the highest reduction of mycelial growth regardless of incubation periods (5 or 10 days). Furthermore, results of cell-free bacterial filtrates showed a significant inhibitory effect on spores germination of M. fructigena, ranging from 10 to 65% (Figure 4). Interestingly, the highest inhibition rates of spore germination were obtained with both bacterial isolates B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11.

Effect of Bacterial Cell-Free Filtrates on Mycelial Growth and Spore Germination
Inhibition rates of mycelial growth of M. fructigena obtained with bacterial cell-free filtrates were highly significantly different between all the 13 isolates ( Figure 3). e cell-free filtrates of both bacterial isolates B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11 have the highest reduction of mycelial growth regardless of incubation periods (5 or 10 days). Furthermore, results of cell-free bacterial filtrates showed a significant inhibitory effect on spores germination of M. fructigena, ranging from 10 to 65% (Figure 4). Interestingly, the highest inhibition rates of spore germination were obtained with both bacterial isolates B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11.

Effect of Bacterial Cell-Free Filtrates on Mycelial Growth and Spore Germination
Inhibition rates of mycelial growth of M. fructigena obtained with bacterial cell-free filtrates were highly significantly different between all the 13 isolates ( Figure 3). e cell-free filtrates of both bacterial isolates B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11 have the highest reduction of mycelial growth regardless of incubation periods (5 or 10 days). Furthermore, results of cell-free bacterial filtrates showed a significant inhibitory effect on spores germination of M. fructigena, ranging from 10 to 65% (Figure 4). Interestingly, the highest inhibition rates of spore germination were obtained with both bacterial isolates B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11.  period, treatments having the same letter are not significantly different according to the LSD test (p < 0.05).

Lower Pathogen Pressure (1 × 10 3 spores/mL)
Two bacterial isolates (B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11) out of the 13 tested bacteria displayed an antagonistic activity against M. fructigena with 100% inhibition after 5 days. However, their efficiencies decreased slightly after 10 days post-incubation as their disease severity increases from 0-0 to 21.70-29, respectively. However, for other bacterial isolates, the disease severity was decreased with the increase of the incubation period (Table 5). Interestingly, the biocontrol protection obtained with B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11 against brown rot disease was statistically comparable to that of fungicidal product thiophanate-methyl (Table 5).

Higher Pathogen Pressure (1 × 106 spores/mL)
The same trend of protection described previously was observed with bacterial isolates at higher pathogen pressure (1 × 10 6 spores/mL). Statistical analyses of disease severity underlined that there was a highly significant difference between bacterial isolates. Surprisingly, both bacterial isolates B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11allowed higher protection against M. fructigena with fewer disease severities for both incubation periods 5 (0-0%) and 10 days (43.59 and 51.09%). The efficacy of these two isolates was significantly higher (5 days) or comparable (10 days) to that of a commercial fungicidal product (thiophanate-methyl, at 1 ppm) ( Table 5).

Lower Pathogen Pressure (1 × 10 3 spores/mL)
Two bacterial isolates (B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11) out of the 13 tested bacteria displayed an antagonistic activity against M. fructigena with 100% inhibition after 5 days. However, their efficiencies decreased slightly after 10 days post-incubation as their disease severity increases from 0-0 to 21.70-29, respectively. However, for other bacterial isolates, the disease severity was decreased with the increase of the incubation period (Table 5). Interestingly, the biocontrol protection obtained with B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11 against brown rot disease was statistically comparable to that of fungicidal product thiophanate-methyl (Table 5).

Higher Pathogen Pressure (1 × 10 6 spores/mL)
The same trend of protection described previously was observed with bacterial isolates at higher pathogen pressure (1 × 10 6 spores/mL). Statistical analyses of disease severity underlined that there was a highly significant difference between bacterial isolates. Surprisingly, both bacterial isolates B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11allowed higher protection against M. fructigena with fewer disease severities for both incubation periods 5 (0-0%) and 10 days (43.59 and 51.09%). The efficacy of these two isolates was significantly higher (5 days) or comparable (10 days) to that of a commercial fungicidal product (thiophanate-methyl, at 1 ppm) ( Table 5).

Semi-Practical Trials of Selected Bacteria
The efficacy of isolates B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11, displaying the lowest disease severity under in vivo conditions, were tested against brown rot disease in a semi-commercial large-scale trial under conditions used by packinghouses ( Figure 5). Their efficacies were compared to a positive control (thiophanate-methyl). Results underlined that all treatments differed significantly from the untreated fruit controls (p < 0.05). Regardless of the treatment, the incidence of brown rot disease was increased gradually with the increase of the incubation period from 10 to 30 days. At 10 days, no infection was detected in apple fruits treated with the bacterial isolate Pseudomonas sp. B11W11 and the fungicide, while treatment with B. amyloliquefaciens B10W10 recorded an incidence of 5%, which was significantly lower than in untreated control (20%). After 20 days of incubation, the incidence of diseased apple fruits treated with Pseudomonas sp. B11W11, B. amyloliquefaciens B10W10, and fungicidal products were 10.37, 14.54 and 20.74%, respectively). However, in untreated control, the disease incidence was much higher, reaching almost 60%. Outstandingly, when untreated apple fruits were completely infected (100%), the incidence of other treatments was the same as that observed at 20 days post-incubation period. Surprisingly, both bacterial isolates offered significantly greater protection against the brown rot disease than that provided by the fungicidal product at 30 days post-incubation period ( Figure 5). Agronomy 2020, 10, x FOR PEER REVIEW 13 of 19

Discussion
So far, when allowed, chemical fungicides are the principal way to control postharvest diseases. However, to achieve a sustainable lifestyle, the use of bacteria as BCAs for fungal diseases has emerged as one of the most reliable and promising alternatives to synthetic fungicides [9,15,34,35]. In this regard, the present study evaluated the ability of 13 bacteria, isolated from natural soil, to control brown rot diseases on apple fruits.
Our results revealed that most of the tested bacterial isolates displayed a significant inhibition rate (more than 70%) of M. fructigena mycelial growth in dual culture bioassay. These results are in agreement with those previously reported on Monilinia spp. using different species of bacteria [9,36,37]. The results underlined that inhibition rates obtained with VOCs (confrontation in the distance) were significantly lower than those obtained with a direct dual bioassay, which might be explained by the fact that isolates use several modes of action to trigger their antagonistic activities depending on the nature of the pathogen. These results are also consistent with those reported by numerous studies [9,36,38]. Furthermore, cell-free bacterial filtrates demonstrated an inhibitory action both on spore germination and on mycelial growth of M. fructigena under in vitro conditions. Similar results were reported by Zhou et al. [39]. In the same line, Liu et al. [40] have reported an inhibition in vitro conidial germination and mycelium growth of M. fructicola in the presence of cellfree culture filtrates of B. amyloliquefaciens C06, implying that it may contain antifungal compounds.
Bacterial antagonists belonging to Bacillus and Pseudomonas genus are well known to be able to secrete antibiotic substances from the lipopeptide family such as fengycin and bacillomycin, as previously reported in several previous studies [9,28]. Our findings support the hypothesis that the antibiosis was extremely involved in the biocontrol activity of the used bacteria. In accordance, antibiosis has been reported as a major factor for bioactivity of B. subtilis CPA-8 in controlling brown rot disease caused by M. laxa and M. fructicola [37].
Few attempts have been made on the biological control of M. fructigena. Some of them are described by Byrde and Willetts [41], including experiments with Trichoderma viride, which has not

Discussion
So far, when allowed, chemical fungicides are the principal way to control postharvest diseases. However, to achieve a sustainable lifestyle, the use of bacteria as BCAs for fungal diseases has emerged as one of the most reliable and promising alternatives to synthetic fungicides [9,15,34,35]. In this regard, the present study evaluated the ability of 13 bacteria, isolated from natural soil, to control brown rot diseases on apple fruits.
Our results revealed that most of the tested bacterial isolates displayed a significant inhibition rate (more than 70%) of M. fructigena mycelial growth in dual culture bioassay. These results are in agreement with those previously reported on Monilinia spp. using different species of bacteria [9,36,37]. The results underlined that inhibition rates obtained with VOCs (confrontation in the distance) were significantly lower than those obtained with a direct dual bioassay, which might be explained by the fact that isolates use several modes of action to trigger their antagonistic activities depending on the nature of the pathogen. These results are also consistent with those reported by numerous studies [9,36,38]. Furthermore, cell-free bacterial filtrates demonstrated an inhibitory action both on spore germination and on mycelial growth of M. fructigena under in vitro conditions. Similar results were reported by Zhou et al. [39]. In the same line, Liu et al. [40] have reported an inhibition in vitro conidial germination and mycelium growth of M. fructicola in the presence of cell-free culture filtrates of B. amyloliquefaciens C06, implying that it may contain antifungal compounds.
Bacterial antagonists belonging to Bacillus and Pseudomonas genus are well known to be able to secrete antibiotic substances from the lipopeptide family such as fengycin and bacillomycin, as previously reported in several previous studies [9,28]. Our findings support the hypothesis that the antibiosis was extremely involved in the biocontrol activity of the used bacteria. In accordance, antibiosis has been reported as a major factor for bioactivity of B. subtilis CPA-8 in controlling brown rot disease caused by M. laxa and M. fructicola [37].
Few attempts have been made on the biological control of M. fructigena. Some of them are described by Byrde and Willetts [41], including experiments with Trichoderma viride, which has not been prepared for commercial application. Also, Falconi and Mendgen [42] showed that isolates of Aureobasidium pullulans, Epicoccum nigrum, Sordaria fimicola and Trichoderma polysporum, applied individually or in combination, provided the best protection against M. fructigena. Larena et al. [43] have shown that the antibiotic produced by Epicoccum nigrum and Candida sake are effective against postharvest pathogens infecting apple fruits, including Monilinia spp. In addition, the treatment of apple fruits with Pantoea agglomerans strain EPS125 has significantly reduced the brown rot disease (M. laxa) incidence and the lesion diameter it causes. Generally, our results are in line with our previous findings in which several bacterial isolates having antagonistic activities against fire blight disease were shown effective against both pathogens associated with the brown rot disease; M. laxa and M. fructigena [9]. These bacteria were especially P. agglomerans, Alcaligenes faecalis and B. amyloliquefaciens [9]. In this study, two bacterial isolates, Pseudomonas sp. B11W11, B. amyloliquefaciens B10W10 were highly effective in controlling M. fructigena under both in vitro and in vivo conditions.
Our results indicated highly significant differences between bacterial isolates to control the postharvest brown rot disease caused by M. fructigena. This ability to inhibit the pathogen varies depending on their nature, number and efficacy of the involved mechanisms and metabolites produced by each bacterium as well. Indeed, it is known that bacteria can inhibit the fungal growth via antibiosis by secreting antifungal substances [35,44], VOCs such as HCN or via parasitism by the production of lytic enzymes such as chitinase, glucanase, amylase, protease and cellulase [45]. Our results underlined that the majority of the 13 bacterial isolates had the ability to produce at least one to two lytic enzymes. In most cases, these metabolites are associated with mycelial degradation, vacuolation and release of cytoplasmic content [34,46]. The role of cellulase in plant disease control has been demonstrated in several studies [47,48]. Downer et al. [49] proved the direct inhibitory effect of cellulase on the fungus Phytophthora sp. Richter et al. [48] also mentioned the role of microbial cellulase in the lysis of fungal hyphae walls as a possible biological control mechanism adopted by bacteria. For both efficient bacterial isolates, only Pseudomonas sp. B11W11 was positive for the production of cellulase, confirming findings reported by Sindhu et al. [47]. The latter described that Pseudomonas strains were able to produce appreciable amounts of cellulase in culture-free supernatants and inhibited the mycelial growth of both pathogenic fungi Pythium aphanidermatum and Rhizoctonia solani. In addition, our results are in complete agreement with the findings of Durairaj et al. [50] who reported no cellulase activity in the potent bacterial strains B. stratosphericus FW3 and Pseudomonas aeruginosa D4.
Several authors have studied the ability of antagonistic bacteria to produce amylase [9,51]. They found that more than 50% of the tested bacteria are able to degrade starch in vitro, and most of these isolates are belonging to the Bacillus genus [9]. Interestingly, both potent isolates Pseudomonas sp. B11W11 and B. amyloliquefaciens B10W10 were found able to produce amylase. These results corroborated those of Kim et al. [51], who highlighted that B. subtilis APEC70 produces lytic enzymes. In a recent study, Mardanova et al. [52] reported similar findings for both antagonistic bacteria B. subtilis GM2 and B. subtilis GM5. However, the α-amylase was considered in several studies as an enzyme with a minor role in the control of certain fungi [53].
Our study emphasized a higher score of protease-producing bacteria isolates in comparison with other previous studies, notably those of El-Sayed et al. [54] and Dinesh et al. [23], who found 23.66% and 27% respectively as a percentage of protease-producing bacteria strains. The role that can play this enzyme in the biocontrol of fungal diseases was discussed in several studies [9,44,45]. Indeed, Dunne et al. [55] evidenced that the inhibitory activity of Stenotrophomonas maltophilia W81 against Pythium ultimum is mainly due to proteolytic activity. Similarly, Budi et al. [56] showed that this enzyme induces in vitro inhibition of Fusarium oxysporum and Phytophthora parasitica by cellular interactions. Our study highlights that the potent B. amyloliquefaciens B10W10 was positive for producing the protease, whereas Pseudomonas sp. B11W11 was not. This result is in agreement with Mounia et al. [57] findings. The latter found that most of the Bacillus species can produce the protease suggesting the crucial role that this enzyme certainly played in the antagonism mechanisms through degrading cell walls of the fungal pathogen.
The ability of bacterial isolates to produce HCN, a secondary metabolite of which glycine is the precursor, was investigated. Although cyanide is a general metabolic inhibitor, it is synthesized, excreted and metabolized by some organisms, including bacteria, as a way to prevent predation or competition [58]. The result showed that only two bacterial isolates, B. amyloliquefaciens B4W4 and B. siamensis B14W14, have the ability to produce HCN. In the same context, several studies have been carried out, generally leading to unpromising results in terms of the number of HCN-producing strains among all the tested bacterial strains; Samuel and Muthukkaruppan [59] and Gautam [60] reported 0% and 7.14%. These low percentages may be the result of the absence of genes carrying HCN production in most rhizobacteria or to the absence of an adequate precursor [61]. The role of HCN in controlling plant diseases has been demonstrated by Voisard et al. [62]. In our study, with the exception of B. amyloliquefaciens B4W4 and B. siamensis B14W14, none of the tested bacterial isolates were shown positive for HCN production. This result is consistent with Sethi et al. [39] findings who claimed the absence of HCN production in the majority of bacterial strains of the Bacillus genus. In our previous study, among 18 bacterial isolates, only two bacterial isolates P. agglomerans ACBP1 and B. mojavensis SF16, were capable of producing HCN [9].
In accordance with in vitro bioassay, our in vivo (apple fruits) results underlined that both inhibition rates of mycelial growth of M. fructigena and disease severity of brown rot were significantly affected by bacterial isolates and pathogen pressure. Both bacterial isolates, B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11 exhibited the highest inhibition rates and the lowest disease severity regardless of the pathogen pressure. These results are similar to those of Mari et al. [63], who found that the antagonistic effect of Aureobasidium pullulans L1 and L8 against three species of Monilinia (M. laxa, M. fructicola and M. fructigena) was higher in vivo bioassays. The two antagonists were selected after preliminary in vitro tests for their ability to reduce mycelial growth. These yeasts have completely inhibited M. laxa, and M. fructicola rots and reduced M. fructigena infections by 70% and 90%, respectively. In our previous study, Lahlali et al. [9] highlighted that B. amyloliquefaciens SF14 and A. faecalis are the most effective bacteria selected among 18 tested bacterial isolates in controlling the brown rot disease.
The effectiveness of both selected bacteria was found to be comparable or even higher to that of fungicidal thiophanate-methyl (1 ppm) and was confirmed in semi-commercial trails. Both selected bacteria have shown an ability to protect apple fruits from M. fructigena attacks through long storage periods (30 days). Finally, the use of these BCAs to control the brown rot disease can be enhanced by their use in combination with other eco-friendly practices. This includes their use, in particular, in mixtures with additives or fungicides at low concentration, as previously demonstrated by previous findings [64,65].

Conclusions
In conclusion, results from in vitro dual culture bioassay, VOCs (in distance effect) and cell-free bacterial culture filtrates highlight the significant effect of bacterial isolates in inhibiting the mycelial growth of M. fructigena. This includes particularly the two bacteria B. amyloliquefaciens B10W10 and Pseudomonas sp. B11W11. The results obtained from both in vitro and in vivo trials confirmed the ability of these two bacteria to reduce the brown rot disease severity and incidence on apple fruits. The efficacy of these BCAs depends on their abilities to produce lytic enzymes and lipopeptides, as it was successfully demonstrated herein. However, further studies should be performed before moving into the next phase of marketing of these bacteria as biofungicide products.