Bacterial strains and isolates
Two strains of P. syringae pv. syringae (BAT13 and DAPP-PG115) were used in this study. DAPP-PG115 was obtained from the Bacterial Collection of the Plant Protection Unit, Department of Agricultural, Nutritional and Environmental Sciences, University of Perugia, (Italy). BAT13 strain was isolated from blast necrosis from citrus trees (cv. Thompson Navel) in the region of Menzel Bouzelfa (Cap-Bon) and stored in the collection of plant pathology laboratory at ISA-CM (Tunisia). Identification of bacterial strains was performed by biochemical tests (LOPAT and GATTa) and by comparing 16S rRNA gene sequences with the GenBank database using the Basic Alignment Search Tool (BLAST). Pathogenicity of the strain BAT13 was confirmed on 1–2-year-old citrus (cv. Thompson Navel), which inoculated with a 108 CFU ml−1 bacterial suspension and compared to the reference strain DAPP-PG115. Symptoms were characteristic of citrus blast. Necrotic areas were developed and enlarged (unpublished data).
Three bacterial isolates (MBCL2, MBCL3, and FCL2) that proved their antagonistic effect to the pathogen, identified by biochemical test and resembling to Bacillus spp., were obtained from symptomless citrus leaves from orchards located in the region of Takelsa (Tunisia). Antagonistic action of those bacterial isolates was proved by in vitro and in vivo tests.
Garlic storage and extraction
Garlic bulbs of Allium sativum were purchased from the supermarket and stored at 4 °C in the dark until required. Axillary buds from the composite garlic bulb were peeled, cleaned, weighed, and roughly crushed. Garlic juice was obtained by squeezing the macerates mixture, using a sterile cheesecloth. The juice was centrifuged, at 4200 rpm for 10 min in order to separate garlic debris from the liquid and filtrated with a syringe filter (0.22 μm). Garlic extract was either used immediately or stored at 4 °C until use.
Analysis of garlic extract for allicin content
The content of allicin was determined spectrophotometrically (Jenway 7315), by the reaction with the thiol, 4-mercaptopyridine. The garlic extract was incubated with 4-mercaptopyridine (10−4 M) in phosphate buffer (50 mM), EDTA (2 mM, pH 7.2), which results in the formation of a mixed disulfide, 4-allylmercaptothiopyridine, and the consequent shift in absorbance at 324 nm was monitored as described by Miron et al. (2002). The negative control was obtained using the same procedure without garlic extract.
Isolation and identification of the antagonistic bacteria
Isolation of the antagonistic bacteria
During surveys, samples were collected from citrus orchards. Young healthy leaves (10 leaves per plant per orchard were sampled) were taken from different citrus orchards located in Takelsa, Chbika, Menzal Bouzelfa, Sidi Bouali, Bouargoub, Akouda, and El Gobba. The leaves were rinsed with sterile distilled water. Each sample was cut into small pieces (about 2 × 2 mm), and then, the fragments were surface-disinfected with 95% ethanol for 3 min. Pieces of tissues were placed in sterilized water and mechanically crushed in a sterile mortar. Then, serial dilutions were made. A loopful of macerate was streaked onto Petri dishes containing the LB (Luria-Bertani) medium and incubated at 25 °C for 3 days.
Identification of the antagonistic bacteria
Potassium hydroxide test (KOH) 3%
The identification of the antagonists was made by biochemical tests. A rapid method (KOH) for the determination of the Gram reactions of bacteria was carried out as reported by Suslow et al. (1982). The bacterium was aseptically removed from Petri plates with toothpick, placed on glass slide in a drop of 3% KOH solution, and stirred for 10 s using a quick circular motion of hand.
Catalase test
A part of the colony in question was transferred to a microscopic slide using a sterile toothpick and mixed with a drop of H2O2. Production of air bubbles is indicative of catalase activity, whereas no air bubbles indicate a lack of catalase activity.
Oxidase
This test determines the presence of cytochrome C oxidase enzyme. Kovacs (1956) method was used. A single colony from a freshly streaked LB agar plate was picked and applied with a sterile toothpick to the discs impregnated with a reagent: N,N,N′,N′-tetramethyl-p-phenylenediamine. The production of a distinct purple color in 10 s was recorded as a positive result.
Hypersensitive reaction (HR) on tobacco plants
In order to ensure that the antagonistic bacteria are not phytopathogenic agents, a hypersensitivity test was carried out on tobacco leaves (Nicotiana tabacum). The bacterial suspension was spectrophotometrically adjusted to (108 CFU/ml) and was injected into the intercellular space of the leaves using a medical syringe. Controls used in this test were a negative control (sterile distilled water) and positive control (strains DAPP-PG115). The absence of complete collapse of the tissue after 24 h was recorded as negative reaction.
Assay of in vitro antimicrobial activity of antagonistic bacteria against P. syringae
Double layer method
Antagonistic activity towards P. syringae pv. syringae of 21 Bacillus isolates (MBCL2, FCL2, GT1, MBCL3, BKT1, GCI1, HT1, FCL1, BKT2, TCK2, TM2, TCK3, MBCL1, MBT1, TCK1, TM4, TM3, TM1, HT2, FCL3, and GCI) obtained was conducted according to the modified method of Vidaver (1976) and Stonier (1960). For each isolate, a bacterial suspension (108 CFU ml−1) was prepared in sterile distilled water (SDW); 20 μl aliquots were spot-inoculated on LB medium and incubated at 25 °C for 48 h. At the same day as the spot inoculation, two P. syringae pv. syringae strains (BAT13, DAPP-PG115) were streaked onto solid King’s B medium and incubated for 2 days at 25 °C. The antagonistic bacteria were then exposed to chloroform vapor for 30 min, and the plates were left open for 15 min in a flow cabinet. One milliliter of bacterial suspension of the pathogen (108 CFU ml−1) was mixed with 3 ml of LB medium (0.6% agar) at 45 °C. This solution was quickly overlain on plates containing the antagonists. Plates were incubated at 25 °C and checked after 24 to 48 h for the appearance of inhibition haloes surrounding the antagonist spots.
Agar well diffusion method
The ability of the antagonist to produce diffusible metabolites was also tested according to the agar well diffusion assay (AWDA) as reported by Tagg and McGiven (1971). The most potential antagonistic bacterial isolates were transferred individually to 50 ml of Luria-Bertani broth medium (LB broth) in a 250-ml Erlenmeyer flask and incubated by shaking at 200 rpm for 4 days at room temperature (RT). Bacterial suspension (1 ml; 108 CFU ml−1) of two P. syringae pv. syringae strains (BAT13 and DAPP-PG115) was mixed with 3 ml of LB medium (0.6% agar) at 45 °C. This solution was quickly overlain on plates containing LB medium, and wells were then punched in the agar with a sterile steel borer. The potential antagonistic cultures were centrifuged at 15,000 rpm for 30 min to remove cell debris. After centrifugation, 100 μl of each sample was aseptically filtered through a 0.45 μm filter and added into the prepared wells. The plates were then incubated at 25 °C, and inhibition haloes around the wells were measured.
Assay of in vitro antimicrobial activity of garlic extract against P. syringae pv. syringae
Disc diffusion method
The disc diffusion method (Pereira et al. 2006) was used to determine the sensitivity of P. syringae towards garlic extract. One milliliter of bacterial suspension (108 CFU ml−1) of P. syringae pv. syringae strains (BAT13 or DAPP-PG115) was thoroughly mixed with LB medium and poured into sterile Petri dish. Many dilutions of the garlic extract were prepared and placed to establish the proportionality of the relationship between the amount of active substance and diameter of inhibition zone. For this, Whatman filter paper disc (6-mm diameter) was placed on LB agar plates surface and an amount of 20 μL of pure garlic extract (100%), 90, 80, 70, 60, 50, 40, 30, 20, and 10% dilutions containing 18, 16, 14, 12, 10, 8, 6, 4, or 2 μl of garlic extract, respectively, were pipetted onto a stack of filter-paper discs. Undiluted garlic extract was considered as the 100% concentration of the extract. Distilled water was used as negative control. Each sterile disc is impregnated by different concentrations of garlic. Then, plates were incubated overnight at 25 °C.
Agar well diffusion method
LB medium was poured into each sterile Petri dish. One milliliter of bacterial suspension (108 CFU ml−1) of the two P. syringae pv. syringae strains (BAT13, DAPP-PG115) was mixed with 3 ml of LB medium (0.6% agar) at 45 °C (Tagg and McGiven 1971). This solution was quickly overlain on plates containing LB medium, and wells of 6-mm diameter were then punched in the agar with a sterile steel borer. Wells were then punched in the agar with a sterile steel borer. One hundred microliters of garlic extract was aseptically added into the prepared wells. The plates were then incubated at 25 °C, and inhibition haloes around the wells were measured. All experiments were carried out with three replicates and were repeated twice in time.
Assay of in vivo antimicrobial activity of garlic extract and antagonistic bacteria against citrus blast disease development
Selected strains showing the best in vitro antagonistic activity levels against P. syringae were used. In this study, strains MBCL2, MBCL3, FCL2 and undiluted garlic extract were evaluated in vivo. One- and 2-year-old citrus plants of cv. Thompson were used. Plants were kept inside a greenhouse in individual pots filled with a substrate composed of peat and sand (2/3v, 1/3v). Twelve plants for each treatment were used. Citrus plants were wounded at six sites on the stem. Each wound site was inoculated with10 μl of bacterial suspension 108 CFU/ml of the strains BAT13 and DAPP-PG115. Three days after inoculation, 10 μl of sterile distilled water (control), or a suspension of antagonistic bacteria MBCL2, MBCL3, and FCL2, crude garlic extract or copper sulfate (0.5%: 0.05 mg/10 μl SDW) as individual treatment was added to the wounds, which were then covered again by Parafilm M. Measurements of the extend of stem necrosis were taken 10 weeks after inoculation.
Data analysis
Data were subjected to analysis of variance using IBM SPSS Statistics software (version 23). Mean values among treatments were compared by Duncan’s multiple range test at the 5% (P = 0.05) level of significance.