Microorganisms used and preparation of inoculants.

Ten strains were used of Bacillus, isolated, and characterized in earlier works, nine strains with activity in vitro against different pathogenic fungi (^{Sosa et al., 2012}: ^{Mejía et al., 2013}; ^{Ruiz et al., 2016}) and one strain (Bacillus cereus BL18) reported with properties in the growth enhancement in Capsicum annum L. (^{Peña et al., 2016}). The bacterial inoculants were obtained from broths in a nutrient broth^® (CN) stirred for seven days at 200 rpm at 29 °C. The time used was to allow the formation of spores, which was verified with a light microscope. Later, samples were centrifuged at 8,000 x g for 10 min; the cellular package was washed with a 0.9 % NaCl solution, and adjusted to a concentration of 1x10⁸ UFC mL^-1 using a Neubauer camera (^{Luna et al., 2013}). Fusarium solani and F. equiseti strains were previously isolated from C. chinense. The pathogens were reactivated in a potato dextrose agar (PDA) medium. For their multiplication, we used oat flakes sterilized in autoclave; for this, 100 g of flakes were placed in 250 mL flasks and 40 mL potato extract were added (300 g of raw potato in 1 L of boiling water for 15 min). Later, we deposited two discs of 0.5 cm of the mycelium in active growth of every pathogenic fungus and they were incubated in a growth chamber at 30 °C for ten days (^{Bruna, 1991}). This was used as an inoculant for later tests.

Inhibition of mycelial growth in vitro of Fusarium spp. by Bacillus spp.

The biotests were carried out by direct confrontation with the widely used dual planting (^{Li et al., 2011}; ^{Essghaier et al., 2012}; ^{Rios et al., 2016}). In the center of 90 x 15 mm Petri dished with potato dextrose agar (PDA) medium, we placed a disc, 0.5 cm in diameter, of mycelium of the pathogen with 10 d of active growth, and innoculated 6 µL of a 1x10⁸ UFC bacterial solution in four equidistant points around the mycelial growth at a distance of 2 cm. The dishes were incubated at 28 °C for 7 d. The radial growth was measured, and using this data, the percentage of mycelial inhibition was calculated, ttaking the fungal growth in PDA in the absence of bacteria as a control.

Pathogenicity tests for Fusarium equiseti and F. solani in habanero chili.

Habanero chili cv. native seeds were used after being disinfected with 2 % sodium hypochlorite and washed 3 times with sterile distilled water. They were planted in plastic trays with 98 holes with the commercial sterile substrate Cosmopeat^®, sterilized in an autoclave at 120 °C for 15 min. Plantlets were obtained 28 days after germination (DDG) and pathogenicity tests were run on them in order to determine the virulence by action of the pathogenic fungi. Pots were prepared using 32 oz styrofoam cups, with 600 g of substrate of a mixture of the soil K´ancab (regional Mayan name) of the type Luvisol and cow manure (1:1) (^{Soria et al., 2002}). Luvisol soils are reddish brown clay soil and slightly acidic (^{Bautista, 2005}). The substrate was sterilized at 120 °C for 15 min. It was added 10 g of oats with the growth of the pathogens of 1x10⁶ conids per grams of oats, which was quantified by taking a gram of oats with inoculant and adding 9 mL of sterilized distilled water. It was counted with the help of a Neubauer camera, one plantlet was placed in each pot, and in order to allow the pathogen to enter the plant, a 1 cm cut was performed on the apex of the root, in order to determine its virulence (^{Herrera et al., 2011}).

Evaluation of the resistance to wilting by Fusarium spp. in habanero chili plants inoculated with Bacillus spp.

The disinfected habanero chili seeds were planted in plastic trays with 98 holes with the commercial sterile substrate Cosmopeat^®. Fifteen days after germination (DDG) an inoculation was carried out, of 1 mL of bacterial suspension of 1x10⁸ UFC mL^-1 on the base of the stems of the plantlets. The control was applied 1 mL of 0.9 % of NaCl. At 28 DDG, the transplant was carried out in 32 oz styrofooam cups with 600 g os substrate (mixture of Luvisol and cow manure), previously inoculated with 10 g of oats that contained 1x10⁶ conids·g^-1. After 28 and 35 DDG, 3 mL of the bacterial solution were added. Fertilization took place with the formula 125-100-150 of N-P-K (^{Soria et al., 2002}). The experiment was established under macrotunnel-type greenhouse conditions in the Instituto Tecnológico de Conkal (21º 04’ N and 89º 31’ O), with six treatments for each pathogen evaluated, four bacterial strains, a treatment that consisted of no bacterial inoculation, but with the presence of the pathogenic fungi, and a control that consisted in plantlets with no inoculation of pathogens of bacterial strains. It was established in a completely random design with 10 repetitions for each treatment; each pot represented an experimental unit. To associate the resistance induced by the bacterial inoculation, every four days after the appearance of the symptoms, the severity was estimated using a scale of five classes: 1=0 %, 3=10 %, 5=25 %, 7=50 %, and 9=100 % damage (^{CIAT, 1987}). Using the severity data, progress curves of the disease were made, and the model of the Area Under the Disease Progress Curve (AUDPC) was used to calculate the intensity of the disease. On the 28th day after inoculation with the pathogens, the final severity of the disease was calculated with the arcsin instruction; y=arsin (sqrt(y/100)), using the parameter of Y_final (^{Campbell and Madden, 1990}). Finally, using the ^{Abbott procedure (1925)}, the efficiency of the bacterial strains as resistance inducers was validated. Averages were compared when the significance between treatments was determined using Tukey (p≤0.05). To verify result reproducibility, the experiments were repeated in two periods (December, 2014 and February, 2015); for the analysis, the averages of both evaluations were used. The statistical analyses were carried out using the program SAS version 9.3 for Windows (^{SAS Institute Inc. 2010}).

In vitro activity of Bacillus spp. against Fusarium spp.

All strains showed inhibition of the mycelial growth of the pathogens; the percentages of inhibition in F. equiseti were between 21.28 and 71.70 %, and for F. solani, it was 37.4 to 69.92 % (Table 1). The greatest effect was observed with B. subtilis CBMT51 on the growth of F. equiseti, and strain CBRF8 showed the greatest effect against both pathogens with values of over 63 %. The strain B. cereus BL8, reported as a growth enhancer, showed percentages of inhibition significantly lower than the two strains mentioned above. The range of inhibition reported by several authors is broad, percentages of inhibition of 90 % are reported for B. subtilis strains against Fusarium sp. (^{Badía et al., 2011}), and, on the other hand, percentages of inhibition of 29.4 % were reported for F. avenacum (^{Essghaier et al., 2012}), and for B. methylotrophicus and B. amyloliquefaciens against F. oxysporum of 42.0 and 51.5 % (^{Rios et al., 2016}); the latter are within the range obtained in this study. Strains CBMT2 and CBMT51 were the only ones to present halos of inhibition of mycelial growth in confrontation with F. equiseti with halos measuring 3.76 and 6.37 mm, respectively, which correspond to 18.3 and 31.5 % of inhibition (Figure 1). The antagonistic activity of Bacillus is due to several mechanisms, such as competition for the colonization of the rhizosphere (^{Compant et al., 2005}), production of lipopeptides such as Iturin, Fengycin, and Surfactin (^{Kim et al., 2010}), and the production of lytic enzymes (^{Pleban et al., 1997}; ^{Chang et al. 2010}). The strain B. subtilis Pla10 produces antibiotics of the family Iturin A and Surfactin; the purified fractions showed that only Iturin A presented fungicidal activity (^{Ragazzo et al., 2011}). The purified chitinase from Bacillus cereus YQ 308 inhibits the elongation of F. oxysporum and F. solani hyphae, reducing the biomass of fungi in comparison with the control (^{Chang et al., 2003}). According to the results, four Bacillus strains were selected for the wilting control tests in habanero chili plants based on its antifungal activity; the strain CBRF8, which showed the highest percentage of mycelial inhibition in F. solani, the strain CBMT51, which showed the highest percentage of inhibition and halo of inhibition in F. equiseti, strain CBMT2, with antifungal activity against both pathogens and the presence of inhibition halo in F. equiseti, and the strain BL18 was included, with a moderate antifungal activity, although reported as having plant growth enhancing properties.

Table 1. Activity in vitro of Bacillus spp. against Fusarium spp. after seven days of confrontation. 

^○Averages with different letters are statistically different (p≤0.05).

Figure 1. Antagonism in vitro of Bacillus subtilis CBMT51 against Fusarium equiseti after seven days of confrontation, compared to the witness. 

Pathogenicity tests for Fusarium equiseti and F. solani in habanero chili.

The pathogenicity tests for Fusarium equiseti ITCF1 and F. solani ITCF2 inoculated in habanero chili plants, showed virulence, since they induced characteristic symptoms such as chlorosis, flaccidity, and partial defoliation partial rotting of roots and neck, which occasionally rose to 3 to 7 cm above the base of the stem. Despite records of 100 % incidence with both species of fungus, severity was greater with F. solani ITCF2, where it recorded 90.0 %, and lower with F. equiseti ITCF1 with 77.5 %, without representing a statistical difference (Table 2). The symptoms of the disease began to manifest themselves after two weeks of inoculation with the pathogen. These results, in terms of incidence and severity, were similar to those obtained in other studies on Capsicum annuum L. and Solanum lycopersicum L., obtaining an incidence of 100 % and severity of 80 to 100 % with two strains of F. oxysporum (^{Apodaca et al., 2004}). In pathogenicity tests on Thevetia peruviana plants inoculated with Fusarium species, F. solani was found to cause symptoms such as necrosis in the stem, chlorotic leaves, and wilting, determining this species as the most virulent (^{Herrera et al., 2011}), results which agree with those in this study. An analysis of variance was carried out with the Area Under the Disease Progress Curve (AUDPC) model to determine the pathogenicty and aggressiveness of the two fungi, and for F. equiseti and F. solani, 1120.0 and 1201.0 units day^-1 were obtained, respectively. There was no significant difference in the virulence of the fungi (Table 2), although a greater unit can be observed of plants diseased with F. solani. For the control, without the presence of fungi, there were 0.0 diseased units, keeping the plants healthy and vigorous until the end of the experiment.

Table 2. Incidence and area under the disease progress curve (AUDPC) for F. equiseti and F. solani in habanero chili plants 

^○Averages with different letters in each column are statistically different (Tukey 0.05).

Evaluation of the resistance to wilting by Fusarium spp. in habanero chili plants inoculated with Bacillus spp.

The appearance of symptoms in plants inoculated with F. equiseti and F. solani were observed eight days after inoculation. In general, in inoculated plants, F. solani displayed more virulence than F. equiseti when inducing greater severity during the progress of the disease; in control plants, without the inoculation of pathogens, the disease was not present (Figure 2). The analysis of the intensity of the disease showed less AUDPC with the inoculations of the strains CBMT51 and BL18 against F. equiseti, although only BL18 allowed less severity in plants infected with F. solani. At the end of the experiment, the parameter Y_final displayed, with CBMT51, a severity of 39.7 % in F. equiseti, while BL18 against F. solani gave 41.7 %. This helped calculate the higher levels of disease control with 47.8 and 50.9 % efficiency, respectively (Table 3). In trials with B. subtilis CAS15 applications on C. annuum, it reduced the severity of the disease caused by F. oxysporum, with a control efficiency of 56.9 %. Similar to that obtained in this study, this efficiency was attributed to a systemic resistance induction (SRI) (^{Yu et al., 2011}). The SRI has been observed in C. annuum against Phytophthora capsici when B. megaterium was used, when reducing the severity of the disease between 18.3 and 33.3 %, and a control efficiency of 23.1 to 57.7 % (^{Akgül and Mirik, 2008}). In the Cucumis sativus crop with F. oxysporum problems, evaluations with Bacillus B068150, reduced the severity by 49.32 %, which meant a control of 50.68 % (^{Li et al., 2011}). In S. lycopersicum, resistance was conferred against Ralstonia solanacearum when using B. subtilis (CYBS-5) with a control efficiency of over 50 % (^{Chen et al., 2012}). Likewise, in S. lycopersicum L. affected by F. oxysporum, the implementation of Pseudomonas fluorescens significantly reduced the incidence and severity of wilting between 4.86 and 74.49 %, where the preinoculation with bacteria tends to improve the activation of the SRI (^{Yu et al., 2011}). Som rhizobacteria ar capable of producing salicylic acid, responsible for SRI, as shown with P. aeruginosa 7NSK2 against Botrytis cinerea (^{De Meyer and Hofte, 1997}). These attributes can be used for the biological control of pathogens that originate in the soil, which can reduce the use of synthetic pesticides.

Figure 2. Disease progress curves of the severity by wilting induced by Fusarium equiseti (A) and F. solani (B) 28 days after the infection in habanero chili plants innoculated with Bacillus spp. 

Table 3. Area under the disease progress curve (AUDPC), severity and control efficiency for Fusarium equiseti and F. solani in habanero chili plants innoculated with Bacillus spp. 

^○Averages with different letters in each column are statistically different (Tukey P≤ 0.05). DMS = minimum significant difference; Control + with presence of the pathogen; Control - without pathogens or bacterias.