Many species of bacteria have been reported to be able to suppress the growth of several fungal and bacterial pathogens and nematodes through antagonist behavior. Among them are included: Agrobacterium radiobacter, Bacillus subtillis, B. megaterium, B. pumilus, B. cereus, Burkholderia cepacia, Pasteuria penetrans, Pseudomonas aeruginosa, P. fluorescent, P. putida, P. syringae, Serratia marcescens, Serratia sp., and Sternotrophomonas maltophilia[24–28].

In this study, a preliminary experiment was carried out to evaluate in the dual plate assay the ability of four B.g. pv. agaricicola strains and of their respective culture filtrates to inhibit the growth of the phytopathogenic and plant-associated fungi B. cinerea, A. flavus, A. niger, P. digitatum, P. expansum, S. sclerotiorum and P. cactorum. Results showed that Bga strains and their culture filtrates inhibited the growth of the selected fungi. Similar antifungal activities have been reported for Burkholderia spp. [19,29,30]. Although in some of the cases the microbial activity was proposed to be mostly due to the fungal cell wall lytic enzymes [30], other antifungal metabolites such as antibiotics, Fe-chelating siderophores, ammonia and cyanide were also reported [19].

Mycelia growth of target fungi in dual plate assays was significantly different from the control (p < 0.05). Strain ICMP 11096 of B.g. pv. agaricicola revealed the highest significant fungitoxicity against A. flavus, A. niger, B. cinerea, P. expansum, P. cactorum and S. sclerotiorum (Table 1). In contrast, strain Bga ICMP 12220 showed the highest significant antifungal activity against P. digitatum (Table 1).

Bga strain ICMP 11096 has been selected for comparing the fungitoxicity of either whole culture or cell-free culture filtrate according to the percentage of fungal mycelium inhibition (PGI). The culture filtrate of bacterial strain ICMP 11096 inhibited the growth of all tested fungi on PDA agar plate assay (Figure 1). The size of the inhibition zones produced by bacterial cells was significantly higher than that produced by the corresponding culture filtrate. On the other hand, there is no significant difference between filtrate and bacterial cells regarding A. niger and S. sclerotiorum, which achieved the same fungitoxicity percentage.

Bga strains showed chitinase, protease and glucanase activities, forming clear zones on chitin, skim milk and Lichenan media plates ranging from 13 to 25 mm, 25 to 36 mm and 20 to 26 mm, respectively. On the contrary, the selected strains did not show cellulase, amylase, pectinase and polygalacturonase activities. The production of chitinase and glucanase could contribute to the growth inhibition control of the phytopathogenic fungi through degradation of fungal cell walls [31]. Protease enzymes as cell wall lytic enzymes, may play a significant role in degrading the fungal cell wall, since fungal cell wall “skeletal” components are embedded in a protein matrix [23] and proteolytic activity is necessary for the lysis of whole fungal cells.

Strains of Bga produced VOCs which reduced the mycelium growth of F. oxysporum (Figure 2). Strains Bga ICMP11096 and ICMP12220 showed the highest significant reduction of fungal growth compared to other studied strains. Strain Bga ICMP11096 has been selected for studying the effect of its VOCs on the fungal growth rate of F. oxysporum (Figure 3). After four days of incubation, the fungal growth appeared to be almost stopped. Similar results were obtained when the above assays were performed versus R. solani (Figure 4). Strains Bga ICMP11096, ICMP11097 and ICMP 12220 showed the highest significant reduction of fungal growth compared to strain Bga ICMP 12322. In contrast, Kai et al.[32] found that volatiles of Pseudomonas spp., Serratia spp., Stenotrophomonas spp. drastically inhibited the growth of R. solani, while a moderate or no inhibition was observed with the volatile substances of B. cepacia and Staphylococcus epidermidis, respectively.

Also in this case the strain Bga ICMP 11096 was able to block completely the mycelium growth after four days of incubation (Figure 5).

GC-MS analysis of VOCs produced by Bga strain ICMP 11096 indicated the presence of 1-methyl-4-(1-methylethenyl)-cyclohexene, which was detected at retention time 11.61 min and has a molecular weight of 136 (Figure 6). This isolated main volatile compound is a liquid hydrocarbon that can be classified as cyclic terpene. This compound could be responsible for the antifungal activity of Bga strain ICMP 11096 against all studied phytopathogenic fungi.

The phytopathogenic as well as plant-associated fungi B. cinerea, A. flavus, A. niger, P. digitatum, P. expansum, S. sclerotiorum, P. cactorum, F. oxysporum and R. solani, were grown and maintained on potato dextrose agar (PDA) or corn meal agar (CMA). Bga strains ICMP 11096, 11097, 12220 and 12322, isolated from Agaricus bitorquis (Quelet) Saccardo, were obtained from International Collection of Micro-organisms from Plants (ICMP) (Landcare Research, Auckland, New Zealand). Bacterial strains were maintained as lyophils at 4 °C and subcultures were obtained on the medium King Agar B (KB) for 48 h at 25 °C.

The antifungal activity of of Bga strains against fungi was carried out by growing the bacterial strains on KB medium and incubated at 30 °C for 24 h. Then 1 cm² agar disc was cut off with a cork borer and placed onto the surface of fungal plates containing PDA medium using bacterial disc method [33]. The plates were incubated at 25 °C for 4–5 days.

The antifungal activity of cell-free filtrates of Bga against fungi grown on PDA was conducted using the diffusion method [34]. Erlenmeyer flask containing 150 mL of liquid minimal mineral medium (MM) were inoculated with 1.5 mL bacterial suspension containing 10⁸ CFU mL⁻¹ and incubated under rotary shaking (180 rpm) at 25 °C for 7 days [35]. Then, the culture was centrifuged at 20,000× g for 15 min and membrane filtered (Millipore, 0.20 μm), then 10 μL of filtrate was deposited on 14 mL solid PDA medium inoculated with 1 cm² of fungal disc. After 4–5 days of incubation, the diameter of the fungal colonies were scored and measured in mm.

According to the screening test, the agar diffusion method was used for testing the antifungal activity of cell-free culture filtrates for the highest bioactive antagonistic bacterial strain of Bga. The antifungal activity of bacterial cultures and filtrates was determined against all studied phytopathogenic fungi and the fungitoxicity was expressed as percentage of growth inhibition (PGI) and calculated according to Zygadlo et al. formula [36] (Equation 1).

where GC represents the average diameter of fungi grown in PDA (control); GT represents the average diameter of fungi cultivated on the treated PDA dish containing the antagonistic bacteria.

Chitinase and protease activities of Bga strains were determined according to Tahtamouni et al.[37] on plates of KB containing chitin (1%) and skim milk (1%) respectively. Cellulase activity was detected according to the Essghaier et al. method [38] using carboxymethyl cellulose (0.4%). Glucanase activity was detected according to Teather and Wood method [39] using Lichenan (0.2%).

Amylase, pectinase and polygalacturanase activities were detected using soluble starch (1%), pectin (0.5%) and polygalacturanic acid (1%), respectively [40,41]. After incubation at 30 °C for 4–5 days, the plates were flooded with specific staining solutions; congo red 0.03% for chitinase, cellulase and glucanase; lugol solution for amylase; CTAB 2% for pectinase; and, ruthenium red 0.1% for polygalacturanase. The enzymatic activity was taken as evidence by appearance of clear zones around the colonies and their diameters were measured in millimeters.

The assays were carried out by plating 100 μL of bacterial suspension containing 10⁸ CFU mL⁻¹ on the surface of Petri dishes containing 14 mL KB incubated at 25 °C for 24 h, whereas fungi were grown on Petri dishes containing 14 mL PDA incubated at 30 °C for 96 h. The bottom parts of the plates were tightly joined with parafilm and incubated in the darkness at 30 °C for 4–5 days. Fungitoxicity of volatile metabolites was expressed by measuring the diameter of mycelium growth (mm). All treatments of the experiment were carried out twice with three replicates. The highest bioactive antagonistic strain was selected to study the effect of its volatile metabolites on fungal growth rate using the Evmert et al. method [42].

For the production of VOCs, the Bga strains were grown on KB and 100 μL bacterial suspensions of about 10⁸ CFU mL⁻¹ were used to inoculate 25 mL glass vials containing 10 mL KB. Vials were sealed with a plastic cap fitted with a Teflon/silicon septum and then incubated for 5 days at 25 °C.

An SPME fiber coated with 100 μm of nongrafted poly-dimethylsiloxane (PDMS) phase (Supelco 57300-U on a Supelco 57330 support) was conditioned for 1 h at 250 °C in helium stream; after that, the fiber was inserted into the headspace of the glass vial containing the inoculated strain and was maintained there for 20 min at 25 °C. The fiber was then introduced into the injection port of a HP 6890 plus gas chromatograph equipped with a Phenomenex Zebron ZB-5 MS capillary column (30 m × 0.25 mm ID × 0.25 μm film thickness). The HP5973 mass selective detector was used (mass range: 15–800 mAU; scan rate: 1.9 scan/s; EM voltage: 1435), helium at 0.8 mL min⁻¹ was adopted as carrier gas. The injection port, equipped with glass insert (internal diameter 0.75 mm) was splitless at 250 °C. The desorption time of 1.0 min was used. The oven temperature was fixed at 40 °C for 2 min, then increased up to 250 °C (8 °C min⁻¹); finally, it was maintained for 10 min. All the analyses were performed in triplicate. A blank run was performed after each analysis in order to confirm that no residual compound was polluting the fiber or the column. The chromatograms obtained from the total ion current (TIC) were integrated without any correction for coelutions and results were expressed as percent of the total area of peaks. All the peaks were identified from their mass spectra by comparison with spectra in Wiley 6N and NIST98 libraries.

The experimental data of the antifungal activity of diffusible and volatile bioactive substances of Bga strains on the phytopathogenic and plant-associated fungi growth were analyzed statistically by one-way ANOVA, Duncan and Tukey B post hoc test using SPSS statistical software (version 13.0, Prentice Hall: Chicago, IL, USA, 2004).