Introduction

Tomato (Solanum lycopersicum L) is one of the most widely sold crops in the world, and one with the highest commercial value. Mexico is the tenth largest tomato producer worldwide, and the single most important exporter, providing the country with an average of one billion dollars. It has been estimated that 1 out of every 3 red tomatoes is produced in the state of Sinaloa, for a total of 867 thousand tons, out of the total of 2.8 millions produced in Mexico (^{SIAP, 2016}). During postharvest, one of the diseases that appear in this fruit is sour rot, caused by the fungus G. candidum, a common alteration in practically any place in which this fruit is produced, and the temperature is preferably of around 30 °C (^{Ruiz-Martínez et al., 2012}). The damage regularly begins through lesions caused by insects, or mechanically, when handling the product in the commercialization process. There are several alternatives to avoid the deterioration of tomatoes and to preserve their quality for longer time periods, including biological control using antagonistic microorganisms such as yeasts, the different action mechanisms include antibiosis, competition for space and nutrients, and direct interaction (^{Bautista-Baños, 2006}; ^{Khaled & Sivasithamparam, 2006}). Among the antagonistic yeasts are what are called killer yeasts, capable of secreting metabolites of a protein nature and variable molecular weights, named killer toxins, capable of inhibiting other microorganisms by altering cell walls, membranes, or vulnerable cell nuclei (^{Buzzini et al., 2007}). There are several reports on the effectiveness and the potential of killer yeasts for the control of fungi that cause plant deterioration. The fungus Botrytis cinerea, one of the most harmful pathogens in grape and strawberry crops, was inhibited using the partially purified of the strain Pichia membranifaciens CYC 1106 (^{Santos & Marquina, 2004}). This same fungus was later inhibited using strains of Pichia anomala and Debaryomyces hansenii, increasing the biological arsenal with biocontrolling potential over this pathogen (^{Santos et al., 2004}). Strains of Issatcehnkia orientalis, Candida guilliermondii, P. ohmeri and Torulaspora globosa, among others, have been reported to successfully inhibit important plant pathogenic fungi, such as Aspergillus cabonarius, A. niger, Penicillium expansum and Colletotrichum sublineolum in grape, pear, apple, and sorghum crops, respectively (^{Bleve et al., 2006}; ^{Coelho et al., 2009}; ^{Rosa et al., 2010}). In Mexico, ^{Hernández-Montiel et al., (2011)} reported a significant reduction of Geotrichum citri-aurantii in limes during postharvest, using epiphytic D. hansenii yeasts.

Yeasts present an ecologically diverse association with insects; over 200 species of only beetles have been reported, along with 12 of termites, 17 for antes, and other insects such as wood-eaters, homoptera, dipterous, hymenoptera, neuroptera, etc. (^{Rosa & Gábor, 2006}). ^{Carreiro et al., (2002)} reported that 77 out of 99 yeast strains isolated from ants of the genus Atta presented the killer phenotype, including some genuses in which this phenomenon had not been previously reported. To study the applicability of this knowledge in preventing the postharvest damage in tomatoes by G. candidum, the aim was set to evaluate the antagonistic capability of killer yeast strains taken from ants, against isolations of G. candidum obtained from tomato samples with characteristic symptoms of sour rot.

Materials and methods

Results and Discussion

From the ants gathered we obtained 8 yeasts, the killer activity of which was observable by the formation of a halo of inhibition, surrounded by a darker blue edge, generated by the methylene precipitated as an indicator of cell death (Figure 1).

Figure 1. Determination of the killer phenotype of yeasts isolated from ants of the genus Atta. Blue edges indicate the cellular death of vulnerable yeast in the “lawn”. 

From the tomato samples, the strain of G. candidum named MG2 produced the most damage in the tomatoes, and was therefore chosen as the most aggressive, and it was used in the antagonism test on the fruit (Table 1 and Figure 2). The comparison of averages using the Tukey test showed that all yeasts, except M5 and M6, helped reduce the significantly higher damage in comparison to the strain of C. albicans and the control without yeast. Strains M1 and M2 were the ones that allowed the least total damage (mm) on the fruits, although with no significant difference in comparison with M3, M4, M7, and M8 (Table 2). The identification of the 8 killer yeasts obtained using the API 20 C AUX test indicated that they all corresponded to Candida (Pichia) guilliermondii, a species in which killer activity has been known for over 20 years (^{Polonelli et al., 1987}), yet it wasn’t until 1991 that its use as a biocontrol agent was first reported (^{Wisniewski et al., 1991}) and its activity has been reported in several occasions (^{Coelho et al., 2009}; ^{de Lima et al., 2013}; ^{Papon et al., 2013}). The difference between the damage in the presence of the killer yeasts and the treatments without yeast and with C. albicans suggest that the phenotype killer in yeasts M1 - M8 is related to its antagonism with G. candidum. The differences observed in the damage to the fruit allowed by each yeast could be explained by differences in the type of killer toxin secreted, which influences the affinity to the white area of the susceptible fungus and/or in differences in the expression of the toxin, since they are mostly codified in the nuclear genome (^{Buzzini et al., 2007}). On the other hand, it is possible that these strains present, not only killer toxins, but also, in a variable way, additional previously reported action mechanisms in C. guilliermondii, such as the direct attack on hyphae and the competition for nutrients (^{Chanchaichaovivat et al., 2008}; ^{Nantawanit et al., 2010}; ^{Scherm et al., 2003}). This diversity in mechanisms represents a diverse interval of antagonistic capabilities for the killer yeasts with plant pathogenic fungi, with the possibility that strains with little or no antagonistic activity towards a genus or species of fungi can show a high antagonistic activity towards others, which allows wide options of study. The use of insects as a nonconventional source of yeasts is attractive, given the size of the biodiversity related to these organisms (^{Rosa & Gábor, 2006}) and the fact that, at least in some cases, associated yeasts represent a relation in which yeast protects the insect from pathogenic fungi (^{Rodrigues et al., 2009}) and with this comes the possibility that this antagonistic mechanism acts against fungi which can be important as plant pathogens.

Table 1. Selection of the most aggressive G. candidum strain, out of a total of 4 repetitions. Results obtained by the inoculation of the pythopathogen by puncturing the fruits are enlisted. 

^a + low, ++ medium, +++ high, determined visually.

Figure 2. Evaluation of the damages caused by different strains of G. candidum on red tomatoes. The presence of mycelia and softening of fruits can be observed. 

Table 2. Average values of the extension of the damage caused by G. candidum in tomatoes in each treatment. 

¹ Each value represents the average of 6 tomatoes. Values in each row with different letters are statistically different (p<0.05).

Although it is clear that they do not represent a direct competition with the efficiency of chemical fungicides due to the speed at which they act, they are an alternative that would help reduce required doses without increasing the damage to the fruits, resulting in products with less concentrations of chemicals, and therefore a reduction in the risk of toxicity for humans.

Conclusions

Results show that it is possible to isolate yeasts with the killer phenotype from ants of the genus Atta and these yeasts display variable antagonism towards G. candidum. The mere presence of a non-killer yeast is not enough to protect the tomato from the damage caused by G. candidum, making it statistically equal to inoculating no yeast. Quantifying the damage to the fruit by accumulated measurement (mm) of the area affected to the sides and the bottom of the lesion help evaluate the protection conferred by the presence of antagonistic yeasts.