Microorganisms used

Sixteen strains of the Bacilli class were used in this study with reported antifungal activity, previously isolated from soil samples from the Yucatan Peninsula, Mexico (Sosa et al., 2012), which belong to the Instituto Tecnológico de Conkal collection, Yucatan. Bacterias were activated on (NA) nutrient agar (BD Bioxon^®, Becton Dickinson Mexico). Its purity was verified by growing homogeneous colonies, the presence of Gram positive cells and endospores formation. All strains were grown in the same medium until autolysis at 28 °C for 4 d and stored at 4 °C until they were use.

Isolation and identification of Fusarium spp.

Samples of stems and roots were collected from habanero pepper plants with characteristic symptoms of the disease, in field and greenhouse areas (21° 04’ N and 89° 31’ W) of research and production, at the Instituto Tecnológico de Conkal. Sections of 0.5 cm were disinfected with 2 % sodium hypochlorite for 30 s, washed twice with sterile distilled water, seeded in Petri dishes with growth medium potato dextrose agar (PDA) (BD Bioxon^®) and incubated at 28 °C for 72 h, then purified. A preliminarily pathogen identification was done based on their morfotaxonomic characters (^{Nelson et al, 1983}; ^{Barnett and Hunter, 1999}).

DNA extractions were taken from monosporic mycelium cultures in order to confirm the specific identification of the fungi (^{Liu et al., 2002}). For this, the ITS1-5.8S-ITS2 rDNA region of was analyzed; primers were those reported by ^{White et al. (1990}), ITS1 (5’-TCCGTAGGTGAACCTGCGG-3 ‘) and ITS4 (5’-TCCTCCGCTTATTGATATGC-3’), amplification was performed by PCR (polymerase chain reaction). This included an initial denaturation for 2 min at 95 °C, followed by 30 cycles of denaturation at 94 °C for 1 min, alignment at 54 °C for 30 sand a 1 min extension at 72 °C, and a final 5min extension at 72 °C. The amplified products were visualized via electrophoresis on 1 % agarose gel (SIGMA^®, USA) and a 1 kb molecular weight marker (Invitrogen^® USA). The amplified PCR products were sequenced at Macrogen (www.macrogenusa.com). The sequences were compared with the GenBank database from the National Center for Biotechnology Information (http:// blast.ncbi.nlm.gov) with the Basic Local Alignment Search Tool (BLAST) program.

Antagonism of bacterial isolates against Fusarium spp.

Bioassays were done in Petri dishes with PDA medium; in the center of each box, 0.5 cm diameter sections of the phytopathogen mycelium and 6 mL of a 1x10⁷ UFC suspension of the bacterial isolates. This was obtained by scraping with a sterile glass slide, from a bacterial culture in AN, after a 5 d incubation period. The suspension was applied at four equidistant points around the fungus, 2 cm apart. The control consisted of Petri dishes with the fungal pathogen but without bacterial strains. The boxes were incubated at 28 °C. After 7 d the mycelial growth inhibition percent (ICR) was determined with the formula proposed by ^{Ezziyyani et al. (2004}), also the inhibition halo, between the fungal colony and bacterial strains. Conidia inhibition was determined by the selecting the strains showing a percentage of ICR of at least 65 % and showed an inhibition halo greater at 3.0 mm at least against one of the fungal strains.

Conidia germination inhibition by filtered bacterial

Bacterial isolates were cultured in 250 mL Erlenmeyer flasks, with a 100 mL of Luria-Bertani liquid medium^® (LB) (DIBICO^®, Mexico), to which 1 mL of bacterial suspension of 1x10⁷ UFC were added. Flasks with the bacterial cells suspension were maintained in an orbital shaker (MaxQ4450, Thermo Scientific, USA) at 200 rpm for 72 h at 29 °C. The supernatant was then recovered by centrifugation at 8765 x g for 10 min, then filtered through Merck^® filters (Millipore Germany) of 0.45 μm. and a bacteria free filtrate was obtained. A suspension of 1x10⁸ conidia mL^-1 of Fusarium spp. was mixed with the bacteria filtrate (1: 1 v/v). The mixture was seeded in Petri dishes with PDA. Each seed point had a coverslip placed on. The number of germinated conidia after 3, 5, 8 and 10 h was recorded with an optical microscope DM500 (Leica Microsystems, Switzerland) with a 100X objective; values were then converted to percentage. This quantification was suspended when in 100 conidia in random control without filtering, the germ tubes extended over half the length of the size of the fungal cells (^{Mahadtanapuk et al., 2007}; ^{Ruiz-Sanchez et al., 2014}).

Identification of isolated bacterial antagonists

The selected Isolates for germination inhibition assays which had previous had a molecular identification were identified. This was done with the analysis of the 16S rDNA gene sequence, and the Eubac27F and Eubac1492R starters (^{Delong, 1992}). To extract the DNA a Wizard Genomic Purification Kit was used (Promega^®, USA). The reaction conditions included 1X PCR buffer, 1.5 mM MgCl₂, 0.2 mM dNTPs mix, 1.0 mM of each oligonucleotide, 2.0 U Taq DNA polymerase (Invitrogen^® USA) and 1mL of DNA at 200ng, to a final 50mL of reaction mixture volume. PCR amplification (in a TECHNE-312, US thermocycler) included initial denaturation at 94 °C for 3 min, followed by 32 cycles (denaturalization at 94 °C for 1 min, 52 °C alignment for 1 min and a 72 °C extension for 1.45 min), final extension at 72 °C for 7 min and 4 °C final temperature. The amplified products and the sequences received the same procedure done with the fungi.

Detection of genes in isolated bacteria for lipopeptide production

The bacterial isolates to detect genes encoding lipopeptid production were the same as those used in the evaluation of the conidia germination inhibition by bacterial filtered, as these were candidates that could produce secondary metabolites with antifungal activity (^{Athukorala et al.,2009}). In order to detect bamC (bacillomycin D), Fend (fengycin), ituA (iturin A), zmaR (zwittermicin A) and sfp (surfactin) genes, the oligonucleotide pairs reported by ^{Ramarathnam et al. (2007}) and ^{Athukorala et al. (2009)} were used. BACC1 F (GAAGGACACGGCAGAGAGTC) and BACC1 R (CGCTGATGACTGTTCATGCT); FEND1 F (TTTGGCAGCAGGAGAAGTTT) and FEND1 R (GCTGTCCGTTCTGCTTTTTC); ITUD1 F (GATGCGATCTCCTTGGATGT) and ITUD1 R (ATCGTCATGTGCTGCTTGAG); ZWITR2 F (TTGGGAGAATATACAGCTCT) and ZWITR2 R (GACCTTTTGAAATGGGCGTA), SUR3 F (ACAGTATGGAGGCATGGTC) SUR3 R (TTCCGCCACTTTTTCAGTTT) (Invitrogen®, USA). The PCR program BACC1 F / R consisted on a denaturation at 94 °C for 3 min, 35 cycles of denaturation at 94 °C for 1 min, alignment at 60 °C for 30 s, extension at 72 °C for 1.45 min and final extension at 72 °C for 6 min (Ramarathnam et al., 2007); for ITUD1F / R and SUR3 F / R consisted of denaturation at 94 °C for 15 min, 30 denaturation cycles at 94 °C for 30 s, alignment at 60 °C for 30 s, extension at 72 °C for 2 min and a final extension at 72 °C for 10 min (^{Stankovic et al., 2012}.); for FEND1 F/R consisted of denaturation at 94 °C for 3 min, 45 cycles of denaturation at 94 °C for 1 min, alignment at 62 °C for 1 min, extension at 72 °C for 1.45 min and final extension at 72 °C for 6 min (Ramarathnam et al., 2007) and for ZWTR2 F / R consisted of denaturation at 94 °C for 3 min, followed by 30 denaturation cycles at 94 °C for 15 s, alignment at 60 °C for 45 s, extension at 72 °C for 2 min and final extension at 72 °C for 4 min (Basurto et al., 2012).

Data analysis

The experimental design was completely randomized and treatments had four repetitions. An ANOVA was performed on the data, with a previous data transformation to arcsin [y=arcsin percentage (sqrt x/100)], and means were compared with the Tukey test (p£0.05). Data was analyzed with SAS Ver. 9.1 (^{SAS Institute Inc. 2010}).

Isolation and identification of fungal strains

Two Fusarium species were isolated from Capsicum chinense Jacq. plants with necrosis symptoms in stems and roots. One was identified as F. equiseti with ITCF1 registration, which had abundant, aerial cottony white mycelium, with rapid growth and orange color on its underside. Microscopic examination revealed the presence of abundant macroconidia with attenuated apical cells, of one to four cells. Microconidia generally present, usually less than simple cells, oval or kidneyshaped produced in false heads. The other isolated species was identified as F. solani with ITCF2 registration. It had sparse aerial mycelium, relatively slower growth, and cream color on both sides. Microscopic examination revealed the presence of few cylindrical macroconidia, round simple basal and apical cells or single cell; abundant microconidia, generally single cell, and in some cases cylindrical and basal attenuated apical foot shaped cells (^{Nelson et al., 1983}; ^{Escalona et al., 2006}). The analysis of the ITS1-5.8s-ITS2 rDNA region confirmed a 100 % similarity of the ITCF1 strain with F. equiseti (GenBank JQ690081) and in ITCF2 case, the similarity was of 99 % regard F. solani (JQ910159). These are important plant pathogens that cause losses in crop production, such as C. annuum. (^{Martinez et al., 2011};^{Sanzon Gomez et al., 2012}), wheat (Triticum aestivum L.), soybean (Glycine max L.) and pea (Pisum sativum L.) (^{Ivić et al., 2009}; ^{Espinoza et al., 2011}), corn (Zea mays L.) (^{Zainudin et al., 2011}), tomato (Solanum lycopersicum L.) and squash (Cucurbita pepo L.) (^{Montano et al., 2012}; ^{Roberti et al., 2012}).

Bacterial isolates antagonistic activity against F. Equiseti and F. Solani

Fusarium spp. ICR. The analysis of variance for the ICR variable showed significant differences between the 16 bacterial isolates (p≤0.05). ICR intervals in both species of Fusarium were from 2.15 to 71.55 %. In F. equiseti ITCF1 bacterial isolates with CBRF8, CBCK41, CBMT51 and CBRM17 registration were those who induced the highest averages of ICR percentage, with 65.70, 66.15, 71.23 and 71.55 % (p£0.05). In F.solani ITCF2 the CBRF15, CBRF9 and CBRF8 isolates outperformed on average the other bacterial isolates (p≤0.05), given that they caused an ICR of 61.60, 62.67 and 69.16 % (Table 1). The antifungal activity of the bacterial isolates become clear in Fusarium sp. when it was confronted with B.subtilis LSB4 and Bacillus sp. LSB9, as they showed mycelial growth inhibition of 90 and 90.4 % each (^{Badia et al., 2011}). Inhibitions in those studies exceeded our results. However, in other cases, the effectiveness with F. oxysporum was similar to that estimated for isolated CBRF8, CBRM17 and CBMT51 (^{Li et al., 2012}b), or less effective, although it came to similar or different species of the fungi (^{Romano et al., 2013}; ^{Sarti and Miyazaki, 2013}) and even when evaluated on other Fusarium strains. Antagonism response is associated with the synthesis of lipopeptides like iturin, surfactin, fengicyn and bacillomycin (^{Ramarathnam et al., 2007}; ^{Mora et al., 2011}; ^{Berić et al., 2012}) and secondary metabolites that can generate an inner lysis area and a thickening edge in the inhibition zone. This effect was observed in F. equiseti and F. solani. They also affected the radial growth of the fungi, as it happened against Curvularia lunata (^{Basha y Ulaganathan, 2002}; ^{Orberá et al., 2009}).

Presence and size of inhibition halo

The analysis of variance of the inhibition halos showed significant differences between the 16 bacterial isolates (p≤0.05). Halos generated in both Fusarium species were of 1.30 to 6.10 mm. In F. equiseti ITCF1 the isolates CBRM17 and CBMT51 were different to the rest of the isolates (p≤0.05), because they generated the largest inhibition halos (5.72 and 6.10 mm). In F. solani ITCF2 only the CBRF15 isolate generated an inhibition halo of 1.30 mm, but it was not the bacteria with the greatest inhibition of mycelial growth (Table 1). The formation of the inhibition halos in bacterial isolates has been evident in the direct contrast of B. cereus X16 against F. roseum (^{Sadfi et al., 2002}). The growth inhibition halos due to the CBRM17 and CBMT51 isolates in F. equiseti showed whitish appearance, with quadrangular delimitation of the fungus. This activity was also observed in Colletotrichum gloeosporioides, by action of B. subtilis and B. licheniformis (^{Gutierrez-Alonso et al., 2003}). The areas bounded by isolates against fungi are due to antifungal metabolites production, which are diffusible in the culture medium (^{Sadfi et al., 2002}). These antifungal metabolites may be cyclic lipopeptides, as surfactin, iturin A and fengicyn, acting as biocontrol agents against Fusarium (^{Arguelles-Arias et al., 2009}).

Inhibition on conidia germination by filtered bacteria

CBRM17, CBMT51 and CBMT2 isolates met the criteria set for this test. The average inhibition halo of F. equiseti exceeded 3.0 mm. But the CBRF8 isolate was determined because of its ICR activity in both phytopathogenic species. Bacterial filtered inhibited between 42.25 and 100 % conidia germination of F. equiseti, and for F. solani the averages were from 12.75 to 100 % (Table 2). The filtering with the highest antifungal capacity was that of the isolated CBRF8, for it inhibited a 100 % of the conidia germination of both fungus species (Figure 1). The 100 % effectiveness was also observed on Colletotrichum musae, exposed to filtered B. licheniformis, B. amyloliquefaciens and B. subtilis; although, it formed swellings and distortions in the conidia and germ tubes of the fungus. Similar activity was not observed in that study (^{Mahadtanapuk et al., 2007}). CBRF8 isolated filtrates are candidates for antifungal activity control, exceeding by 40 % the effectiveness of filtered B. subtilis applied to Aspergillus flavus conidia (^{Zhang et al., 2010}), and 63 % to those from Bacillus sp. on C. gloesporioides conidia (^{Gutierrez-Alonso et al., 2003}). The CBRF8 bacterial isolate did not showed inhibition halos in the tested fungal strains, but the antifungal capacity of the filtered may be associated to the production of proteins with enzymatic or toxic activity, which together cause somatic structural changes on the fungi and in the conidia viability (^{Basurto-Cadena et al., 2010}). These can affect the overall growth of the fungi or cause only morphological changes in the mycelium or in its resistance and reproduction structures (^{Carissimi et al., 2009}).

Figure 1 Conidia germination inhibition of Fusarium equiseti by the CBRF8 isolated filtering after 24 h of exposure: A, treated conidia; B, untreated control (100X). 

Identification of two isolated bacterial antagonists

CBMT2 and CBMT51 isolates were identified as B. subtilis (Ruiz t al., 2016); so, in our study the CBRM17 and CBRF8 isolates were analyzed, based on the 16S rDNA gene sequence. The amplified PCR product was about 1500 bp, and had a 94 to 99 % similarity with the CBRF8 isolated with Bacillus subtilis (GQ214132.1) and CBRM17 with Paenibacillus sp. (KJ948328.1). Although these isolates are of different genera, they belong to the same Bacilli class. There is documented diversity of Bacillus antagonists, and B. subtilis is the one that has the most recognized antifungal properties against Fusarium (^{Guillen-Cruz et al., 2006}; ^{Badía et al., 2011}; ^{Li et al., 2012}a). Also Paenibacillus polymyxa, P. peoriae, P. brasilensis, P. alginolyticus and P. favisporus show antifungal activity against F. oxysporum (^{Li et al., 2012}b; ^{Sato et al., 2014}).

Detection of genes encoding antifungal lipopeptides biosynthesis

Amplified specific products were only observed in the CBRF8 bacterial isolate with the expected size for the bamC, ituA and sfp gene fragments of 875, 675 and 500 bp, each (^{Ramarathnam et al., 2007}; ^{Berić et al., 2012}; ^{Stanković et al., 2012}.) (Figure 2). In the CMBR17, CBMT51 and CBMT2 isolates there was no detection of these fragments. The detected genes in the isolated CBRF8 coded bacillomycin D production, iturin A and surfactin. The presence of more than one gene for lipopeptides synthesis in the B. subtilis CBRF8 isolate, is a characteristic of the isolates of this genere, in which there has been obtained 167 isolates having the bacillomycin D gene, 111 the surfactin gene and 79 the iturin gene (Stankovic et al., 2012). In same and different species the bacillomicyn genes presence has been detected, as in B. cereus DFE4, B amyloliquefaciens DEF16 and BS6 and B. subtilis 49, the fengycin gene in B. subtilis DFH08 and 49; although there was no amplification for this gene, its synthesis is associated with 60 % growth inhibition of F. graminearum and Sclerotinia sclerotiorum (Ramarathnam et al., 2007). This species presents iturin C and D genes and bacillomicin C and AB, related to radial growth inhibition, inhibition halos and inhibition of conidial germination in F. oxysporum and F. solani (^{Chung et al., 2008}). These mechanisms of lipopeptide synthesis genes with antifungal metabolites or compounds of various peptide and lipopeptide nature, such as iturins, fengycins and surfactins, can interfere with the antifungal activity of B. subtilis against phytopathogens originated in the ground (^{Cazorla et al., 2007}).

Figure 2 PCR amplification of the fragments of genes for lipopeptides biosynthesis from CBRF8 isolated, with the specific oligonucleotides. Lane M, molecular weight marker (1 kb DNA Ladder; Invitrogen ^® ); lane 1, ITUD1; lane 2, FEND1; lane 3, BACC1; lane 4, ZWITR2; lane 5 SUR3; lane 6 negative control.