Antagonistic effect of Trichoderma harzianum VSL291 on phytopathogenic fungi isolated from cocoa (Theobroma cacao L.) fruits

In this study we evaluated the antagonism in vitro of Trichoderma harzianum strain VSL291 against 18 pathogens of cocoa fruits in dual culture. T. harzianum VSL291 inhibited the growth of the phytopathogenic fungi tested between 10.54 and 85.43%. The mycoparasitism of Moniliophthora roreri by T. harzianum VSL291 was studied by light and scanning electron microscopy. T. harzianum VSL291 hyphae grew in parallel with the hyphae of M. roreri and in some places these were united with the hyphae of the cocoa pathogen through small structures like apresorious that tangled in the pathogenic fungus preventing its growth. T. harzianum VSL291 produced lytic enzymes: β-1,3-glucanases, chitinases, proteases, xylanases and lipases, when grown in minimal medium, with fungal cell walls as the sole carbon source. The highest proteolytic activities detected in T. harzianum VSL291 broth with M. roreri, Penicillium expansum and Byssochlamys spectabilis cell walls appear to be associated with increased activities of -1,3 glucanases, chitinases, lipases, proteases and xylanases and biocontrol index derived from the experiments of confrontation. These results suggest that proteolytic enzymes according to their degree of induction could participate in the antagonistic effect of T. harzianum VSL291 against the fungi tested.


INTRODUCTION
In the year 2010, the world wide cocoa was estimated in 3.7 millions of ton, of which Africa was the main producer with 2.51 millions of ton, Asia and Oceania 0.66 millions of ton and Latin America and the Caribbean with 0.51 millions of ton (FAO, 2004). However, in cocoa plantations a lot of wastes are made that generally remains scattered into the plantation and causing the *Corresponding author. E-mail: mario.ramirez.lepe@gmail.com. Tel/Fax: 229-9345701.
Abbreviations: BCI, Biocontrol index; PDA, potato dextrose agar; SD, standard deviation phytopathogen fungi propagation that affect the cacao plant and play an important role in the destruction of the important natural resources of the cocoa industry. In the last decades, the worldwide cocoa production was seriously affected by diseases caused by phytopathogenic fungi. For example, in tropical America the more important were the frosty pod rot caused by Moniliophthora roreri (Phillips-Mora et al., 2006) and the witches broom caused by Crinipellis perniciosa (de Marco et al., 2003). On the other hand, diseases caused by Phytophthora strains produce cocoa losses worldwide between 45 to 100% of the production (Djocgoue et al., 2010). The cocoa plants cultivated in the Chontalpa sub region located in the state of Tabasco, Mexico, are affected by frosty pod rot disease (Phillips-Mora et al., 2006), which cause losses higher than 80% (Phillips- Mora and Wilkinson, 2007). Many chemical products applied to control this disease are very effective in most of the case, but caused serious damage to the environment, soil, and humans (Lee et al., 2004). Other techniques of control are the plant management with cultural practices and the genetic resistance (Howell, 2003), but these techniques only reduce the incidence of fungal pathogens into the plantations.
Trichoderma strains are antagonistic to some phytopathogenic fungi because they have the ability to suppress the diseases they cause (Harman, 2006). Trichoderma uses several biocontrol mechanisms such as mycoparasitism, antibiosis, and competition for space and nutrients, and is also able to promote plant growth and development, and induce the defense response of plants (Shoresh et al., 2010). It has also has been demonstrated that it has the ability to produce lytic enzymes that can act in a synergistic way increasing its antagonist action (Benitez et al., 2004). Several Trichoderma species have been extensively studied for their biocontrol potential of diseases in different crops (Harman et al., 2004). However, despite the importance of M. roreri in cocoa farming, little is known about the interaction of Trichoderma-cocoa phytopathogenic fungi (Samuels et al., 2006). The aim of this study was therefore to study the antagonistic effect of T. harzianum VSL291 on fungal plant pathogens isolated from cacao fruits with symptoms of disease.

Fungal strains
The microbial strains used in this study included the following phytopathogenic fungi: M. roreri, Phytophthora megasperma, P. capsici, Colletotrichum gloeosporioides, Fusarium solani, F. coeruleum, F. verticillioides, Corynespora cassiicola, Cochliobolus lunatus, C. hawaiiensis, Cladosporium cladosporioides, Byssochlamys spectabilis, B. nivea, Penicillium chrysogenum, P. expansum, Rhizopus oryzae, Neurospora crassa and Aspergillus niger. The strains were isolated previously from diseased tissue of cocoa fruits with symptoms of frosty pod rot and pod rot disease, from Huimanguillo, Tabasco State in Mexico. The antagonistic fungi T. harzianum strain VSL291 was isolated from the soil cultivated with Agave tequilana cv. 'Azul' in the State of Jalisco, Mexico (Sánchez and Rebolledo, 2010). All were obtained from the culture collection of the Genetic Laboratory of the Instituto Tecnológico de Veracruz.

Confrontation experiments
The confrontation experiments of the interactions between eighteen phytopathogenic fungi and the antagonistic T. harzianum strain VSL291 was evaluated by using the technique described by Szekeres et al. (2006). Briefly, in Petri dishes with PDA medium, three-day-old T. harzianum mycelia discs of 5 mm in diameter were put at equidistant points and each fungal pathogen was left to confront. The Petri dishes were incubated at 25°C in darkness and then the antagonist activities were recorded on day 7 of incubation and digital images were taken at a distance of 18 cm with a Cybershot DSC-P72. The percentage of inhibition of the growth of the pathogen fungi was calculated using the biocontrol index (BCI) according to the formula: BCI = [A / B] × 100 (Szekeres et al., 2006); where A is the area of the colony of T. harzianum and B is the total area occupied by colonies of T. harzianum and each pathogen fungus. The Image software (http://www.ansci.wisc.edu/equine/parrish/index.html) was used to calculate the area for the BCI. Data of inhibition were compared among the phytopathogenic fungi by ANOVA, followed by Tukey´s tests as warranted, and inhibition areas values were compared among the phytopathogenic strains using 95% confidence intervals. All statistical analyses were conducted with the Statisticx 9.0 software (Analytical Software, 2008).

Mycoparasitic assays
In order to complement the macroscopic observations already observed, microscopic descriptions were made on the mycoparasitism exerted by T. harzianum on the fungal pathogen M. roreri known to be responsible for the greatest number of losses in the cocoa plantations in Mexico. To observe the region of interaction between Trichoderma and M. roreri, observations were made by scanning electron microscopy (SEM). Electron micrographs were taken at the Institute of Ecology (INECOL, Xalapa, Veracruz, Mexico). The samples analyzed came from the interaction zone of confrontation between pathogen and antagonist in the potato dextrose agar (PDA) medium. The processing of samples was done by the following method: agar cuts of 2 × 2 mm were prepared from the zone of interaction between T. harzianum and M. roreri. The samples were immersed in an aqueous solution of 1% agar and set at 4°C for 2 h by immersion in a dissolution consisting of glutaraldehyde (25%) to 3% in sodium catodylate buffer, 0.1M at pH 7.2, followed by three washes with the same buffer for 30 min in the dark, then fixed by immersion in osmium tetroxide 1% for 2 h at 4°C in the dark. Dehydration was carried out in stages of 15 min with increasing ethanol dissolutions of 30, 50, 70, 90 and 100%, and dried by the critical point method (ethanol/CO2 liquid). The cuts were mounted on a pedestal with graphite conductive paint and coated with gold by evaporation method and sputtering. The test was done with a SEM JEOL, JSM-5600LV scanning electron microscopy (Bozzola, 2007).
Also, enzymatic activities were measured every 24 h. The β-1,3glucanase activity was measured using the method described by Miller et al. (1959). Using laminarin as substrate, one unit of enzyme activity was defined as the amount of enzyme that produced 1 mmol of reduced sugar (using glucose as a standard) per ml per minute. Xylanase activity was determined according to Rawashdeh et al. (2005). Using xylan as substrate, one unit of enzyme activity was defined as the amount of enzyme required to release 1 μmol of reducing sugar (using xylose as a standard) per ml per min. Chitinase activity was determined according to the method described by Monreal and Reese (1969). Using colloidal chitin as substrate, one unit of enzyme activity was defined as the amount of enzyme required to release 1 μmol of reducing sugar (using N-acetylglucosamine as a standard) per ml per min. Protease activity was measured using the method described by Kunitz (1946). The enzyme activity was defined as the amount of enzyme required to release 1 μg of tyrosine per ml. per min. Lipase activity was measured using the method described by Nawani et al. (1998). The enzyme activity was defined as the amount of enzyme that produced 1 mg of p-nitrophenol per ml per ml. The protein content was estimated according to Bradford (1976). Statistical analyses of the data were done as aforementioned.

Confrontation experiments
On the digital images, the colonies of T. harzianum VSL291 strain as well as the total areas occupied by the colonies of both Trichoderma and the cocoa tree pathogen were drawn around and measured by the ImageJ software with the use of freehand selection tool. During the analysis, the scale was set to 28.346 pixels per cm; accordingly the unit of the calculated areas was cm 2 . The areas of Trichoderma VSL291 colonies were measured daily for one month and the changes of the occupied areas were followed (data not shown). T. harzianum VSL291 have control over most of the phytopathogenic fungi strains tested. An initial rapid increase of the Trichoderma VSL291 colonies was observed in most cases at the 5 to 7 days, while the areas did not change considerably during the following period and remained approximately the same from the seventh day to the end of the investigation. A zone of progressive inhibition produced by T. harzianum VSL291 against, C. lunatus, C. gloeosporioides, C. cassiicola, C. cladosporioides, F. verticillioides and M. roreri was observed, while R. oryzae, P. megasperma, and P. capsici produced less inhibition.
The digital images taken on the seventh day were used to calculate the BCIs values. The average and the standard deviation (SD) values were calculated from the three replicate measurements for both areas; T. harzianum area and T. harzianum + phytopathogenic area (Table 1). The higher effect of T. harzianum on the growth of the phytopathogen fungi were obtained with P. chrysogenum, C. gloeosporioides, P. expansum, B. nivea, and C. lunatus. The percentage of inhibition ranged from 76.37 to 85.43%. On the other hand, the lower inhibition effect was observed with N. crassa, P. capsici, P. megasperma, and R. oryzae, which ranged from 10.54 to 35.06%. The percentage of inhibition that Cuervo-Parra et al. 10659 we found for a microorganism of the genus Penicillium (P. chrysogenum) was higher than that previously reported for the biocontrol of P. digitatum by T. harzianum. However, for P. expansum and B. nivea, our results were lower than those reported for the same microorganisms (Guédez et al., 2009). Quiroz et al. (2008) found that the highest inhibition of the growth of Penicillium sp. and Fusarium spp. were obtained when the confrontation experiments were carried out with the Trichoderma sp., strains RP-12b and ST-2 whose percentage of inhibition ranged from 70 to 100%. Those percentages of inhibition agreed with our results. For the microorganisms of the genus Phytophthora however, our results were lower to that obtained by other authors for biocontrol with T. harzianum strains (Aryantha and Guest, 2006;Villegas and Castaño, 1999). For the genus Fusarium, our results of BCI values agreed with that obtained by Suárez et al. (2008).

Observation of mycoparasitism by scanning electron microscope
The mycoparasitism was analyzed by observing the preparations by scanning electron microscope. It was possible to observe the morphological structure and distribution of the hyphae of M. roreri in pure culture ( Figure 1a) and the changes that these structures experienced when the pathogen fungi was confronted in dual culture with T. harzianum (Figure 1b and c). During the confrontation period in dual culture, the hyphae of T. harzianum grew on the hyphae of M. roreri, causing morphological deformations and disorganization in the structure of their cell wall so that its appearance becomes rough, probably due to the secretion of antifungal substances (enzymes and antibiotics) by T. harzianum (Figure 1b and 1c). The disintegration of mycelial walls resulted in the total destruction of the colony of M. roreri. Moreover, optical microscopy examination of the hyphae of T. harzianum in the interaction zone, showed that hyphal interactions exist, which would demonstrate a parasitic behavior of strain T. harzianum VSL291 on M. roreri. It envisioned the growth surrounding the hyphae of T. harzianum on M. roreri (Figure 1d), wrapping them either loosely or tightly. Also, it was observed that hyphae of T. harzianum normally grew in parallel to those of M. roreri (data not shown) and that at certain intervals were connected to them with small branches like apresorious (Figure 1e). These results demonstrate the particular parasitic ability of the antagonist that finally is able to inhibit the growth of the pathogen. These morphological changes in the structure of the phytopathogenic fungi caused by T. harzianum were similar to those observed with other phytopathogenic fungi. Benhamou et al. (1999) reported some events of the mycoparasitism between Pythium oligrandrum and  Figure 1); b Area T + P: total area occupied by the colonies of Trichoderma and pathogen fungi (yellow line in Figure 2); c Biocontrol index at day 7. Different letter within the column indicate significant differences (P  0.05, ANOVA and Tukey´s tests).   other phytopathogenic oomycetes (Rhizoctonia solani, F. oxysporum, P. megasperma, and P. ultimum). The damages made on the different fungal structures of the pathogenic fungus were related with the presence of increased sizes of cells derivates from the disorganization of the cytoplasm, retraction and rupture of the plasmatic membrane and the alteration and distortion of the cell wall in the place of penetration of the antagonist. All these action therefore trigger the massive colonization of the pathogen and causes cellular lyses.
Previous studies have shown that T. harzianum chitinases and glucanases activities were induced when the cultures were supplemented with cellular walls of Sclerotium rolfsii (Elad et al., 1982), F. oxysporum, R. solani (Sivan and Chet, 1986), Botrytis cinerea (Schirmbock et al., 1994) and C. perniciosa (de Marco et al., 2003). On the other hand, Suárez et al. (2005) using the strain CECT 2413 of T. harzianum grown with cell walls of B. cinerea, R. solani and P. ultimum as a sole carbon source, found that T. harzianum proteomic response varies both qualitatively and quantitatively on the different fungal cell walls, suggesting that T. harzianum is able to modify the production of these proteins according to the fungal host. Our results are consistent with the results of these groups, T. harzianum VSL291 induced the synthesis of enzymes according to the cell wall of fungi tested. Our results of -1, 3glucanase activities of cell wall tested are similar to those obtained by Sivan and Chet (1989) with mycelia of R. solani and by de la Cruz et al. (1995). This group reported activities of 0.700 to 1.0 mU/µg protein for cell walls of B. cinerea, Gibberella fujikuroi, R. solani, Phytophtora citrophthora, and Saccharomyces cerevisiae using a strain of T. harzianum. However, Vazquez-Garcidueñas et al. (1998) reported -1, 3-glucanase specific activities from 1 to 27 mU/µg protein in 48-h T. harzianum broth with fungal cell walls of Macrophthalmus rouxii, Neurospora crassa, R. solani, and S. cerevisiae.
On the other hand, Küçük, and Kivanç (2008) obtained chitinolytic activities values similar to those obtained by us, using T. harzianum strains grown in liquid cultures containing Gibberella zeae and Aspergillus ustus cell walls as sole carbon source. The values obtained in our study for A. niger and F. verticillioides were lower to that reported by this group. In another study, Rey et al. (2000) using wild and modified strains of T. harzianum grown in cell walls of S. cerevisiae and B. cinerea as sole carbon source obtained higher chitinolytic than that obtained in our study. Xylanase and lipase activities detected in this study are low probably because the concentration of xylan and lipids in the cell walls are zero or very low and hence the induction is low. To the best of our knowledge there are no xylanase and lipase activities reported by Trichoderma using fungal cell walls as sole carbon source, probably because the concentrations of xylan or lipids in the cell walls are very low. Levels of xylanase activity of our work ranged from 0.08 to 1.82 mU/mg protein, significantly lower than those reported from xylan as carbon source ranging from 206 to 24 400 mU/mg protein (Stricker et al., 2006;Seiboth et al., 2003). With regard to lipase, values obtained in this study ranged from 0.01 to 0.60 mU/mg protein, significantly lower than that reported by Kashmiri et al. (2006) from olive oil as carbon source. The results of the low levels of induction of lipolytic and xylanase enzymes by T. harzianum VSL291 are congruent because these enzymes are inducible by xylan and lipid concentration and those in cell walls of fungi are low (Sentandreu et al., 2004).
More also, we chose the cell walls of fungi as a carbon source and measured products released to determine if the amount of products released by the hydrolysis of cell walls could be related to its BIC. We found no correlation between BCI and the degree of enzyme induction (data no shown), probably because in confrontation experiments there are other control mechanisms such as production of antibiotics (Tijerino et al., 2011) and competition for nutrients (Elad et al, 2000). However, our results show that the cell walls of strains M. roreri, P. expansum, B. spectabilis and C. hawaiiensis have the highest proteolytic activity in the T. harzianum broth.
Interestingly, M. roreri, P. expansum, and B. spectabilis also have high BCI (Table 1).These results therefore show that proteolytic activity of T. harzianum VSL291 could play an important role in mycoparasitism. We suggest that T. harzianum VSL291 proteases could expose other components of the fungi cell wall, inducing the expression of other enzymes and consequently increasing the antifungal activity. In relation to fungi cell walls that had low levels of proteolytic activity we suggest that the affinity for the substrate and the degree of exposure to cell wall proteins of T. harzianum VSL291 proteases could have influenced these low activity.
These results are consistent with other reports. Suarez et al. (2004) have shown that protease PRA1 from T. harzianum CECT2413 has different degrees of expression according to the substrate used and has additive or synergistic effects with other proteins produced during the antagonistic activity. Howell et al. (2003) have attributed in part the action of proteases produced by Trichoderma strains to inactivate hydrolytic enzymes produced by pathogens. Benitez el al. (2004) also reported that alkaline protease Prb1 from T. harzianum IMI 206040 plays an important role in biological control and Prb1 transformants showed an increase of up to fivefold in the biocontrol efficiency of Trichoderma strains against R. solani. On the other hand Sivan and Chet (1989) reported that the cell walls of the Fusarium species contain more proteins than the cells of other fungi and this makes it difficult for the degradation of their cell wall. In general, our results therefore show that the cell walls that induced in T. harzianum VSL291 higher levels of proteolytic activity also induced a higher activity of other enzyme activities tested.