Introduction

Apple fruit production in 2016 worldwide was 89.3 million ton. Mexico contributed with 0.716 million ton in the same year (^{FAO, 2016}). Within Mexico, Chihuahua State is the number one producer of apples with an average production of 0.586 million of ton per year (^{SIAP, 2017}). Golden Delicious is the most important cultivar grown in Chihuahua State. One of the main problems that occurs during the postharvest storage of apple fruit is the damage due to Botrytis cinerea (^{Yu et al., 2007, 339}; ^{Calvo et al., 2007, 251}; ^{Williamson et al., 2007, 561}; ^{Xiao and Kim, 2008, 1}; ^{Li et al., 2011, 151}; ^{Guerrero-Prieto et al., 2011, 91}). , which is still mainly controlled by using chemical synthetic fungicides (^{Chand-Goyal and Spotts, 1996, 253}; ^{Droby et al., 2009, 137, 138}; ^{Eshel et al., 2009, 48}; ^{Quaglia 2011, 307}; ^{Feliziani et al., 2013, 133}; ^{Wisniewski et al., 2016, 3}; ^{Sandoval-Flores et al., 2018, 207}) that contribute to subsequent human health and environmental risks and the development of microorganism resistance (^{Mari et al., 2014, 1}). Botrytis cinerea remains quiescent in the host and grows at different temperatures, 0°, 4°, 12°, 15°, 25° and 28°C (^{Fernandez et al., 2014, 541, 542}) and starts damage at different number of days, depending on the storage conditions and the host physiology (^{Williamson et al., 2007, 561}). The damage caused by Botrytis cinerea in different produce is of a significative economical importance worldwide (^{Williamson et al., 2007, 561}). In developing countries, like Mexico and the apple growing area of Chihuahua, Mexico, (^{Guerrero-Prieto et al., 2013, 76}) reports a loss of postharvest fruit of 50% of damage due microorganisms, including Botrytis cinerea on postharvest apples. The damage caused on apple fruit because of Botrytis cinerea starts in the physically damaged areas of the fruit, showing symptoms with brown color of rotten soft pulp tissue, which increases in size as the fruit stays longer in the cold storage (^{Williamson et al., 2007, 562, 563}; ^{Xiao and Kim, 2008, 2, 3, 4}). The use of yeast as a biocontrol agent is an alternative to the use of fungicides (^{Chand-Goyal and Spotts, 1996, 253}; ^{Spotts et al., 2002, 252}; ^{Droby et al., 2009, 138}; ^{Eshel et al., 2009, 48}; ^{Quaglia 2011, 308}; ^{Sowndhararajan, et al., 2013, 1492}; ^{Mari et al., 2014, 2}; ^{Droby et al., 2016, 22, 23}; ^{Sui et al., 2016, 34}; ^{Wisniewski et al., 2016, 4}). In the apple growing area of Chihuahua, Mexico, several of the most commonly used fungicides are; Thiabendazoles, Benomyl, and Captan (^{Ramírez-Legarreta and Jacobo-Cuéllar, 2002, 172}) and can be applied to the fruit by a dip in water, which is used for a bath before it is either sorted to be packed and before being cold stored (^{Sastre et al., 1999, 182}). A study like the present one, will be helpful to the apple growers in the Mexico's number one apple producing area, since biological control agents, like Candida oleophila, work better when obtained from the same fruit, as in the present work (^{Guerrero-Prieto et al., 2004, 223}). Candida oleophila (Co) strains used in the present research work were isolated from apple fruit grown in the Chihuahua apple production area by (^{Guerrero-Prieto et al., 2004, 223}) and this is the reason for using this yeast, which is common to apple fruit and other fruits in the world (^{Droby et al., 2009, 138}). Co can grow at different temperatures and is able to grow at 0° C, which is a key characteristic since apple fruit is stored at 0° C and can control pathogens such as Botrytis cinerea (^{Guerrero-Prieto et al., 2004, 225}; ^{2011, 97}). One of the modes of action used by Co to control Botytis cinerea is nutrient competition (^{Guerrero-Prieto et al., 2011, 97}) and production of exo-ß,1-3 glucanase production (^{Guerrero-Prieto et al., 2014, 431}). With the hypothesis that the combination of Candida oleophila and the synthetic fungicides, will allow to reduce the use of synthetic fungicides, the objectives for this research work were, to explore the resistance of the three Candida oleophila (Co) strains to the synthetic fungicides, in vitro, by measuring colony growth, alone and in combination with the fungicides being tested, and to evaluate the control in vivo of Botrytis cinerea, in postharvest Golden Delicious apple using the strains, to improve control efficiency and reduce fungicide use.

Method

Candida oleophila strains L06, L07 smooth and L07 rugose used in this research, as well as Botrytis cinerea, were obtained from “Temperate zone microorganism’s collection” at Centro de Investigación en Alimentación y Desarrollo, A. C. (CIAD, A. C. www.ciad.mx) Cuauhtémoc Unit, Chih. México. Both, Candida oleophila and Botrytis cinerea were not checked morphologically because they came from the “Temperate zone microorganism’s collection” and this is a guarantee for the authenticity of the microorganisms used. Inoculum concentrations were determined based on previous works from (^{Guerrero-Prieto et al., 2011, 92}) and it is common to use higher inoculum concentrations than the natural concentration found in the original yeast and fungus concentration in the fruit. Concentrations used for each one of the three Candida oleophila strains (L06, L07 smooth and L07 rugose) were 1.1x10⁹ cfu/mL (colony forming units per ml) and for Botrytis cinerea 1x10⁶ conidia/ml, in both cases using a Neubauer chamber for determining inoculum concentration (^{Renping et al., 2011, 152}). In vitro Candida oleophila colony growth. Growing media used was PDA (potato, dextrose, agar, 39 g/lt) and the fungicides were mixed with the PDA. Once the Petri dishes with PDA plus the fungicide were inoculated, they were incubated at 22°C for one week. Based on fungicide manufacturer recommendations, three doses for each commercial fungicide were used; low (lower than recommended), medium (manufacturer´s recommendation) and high (higher than recommended), as follows; Cyprodinil+Fludioxonil (Switch^® 62.5 WG Syngenta), 0.75, 1.0, 1.5 g/l; Benomyl (Benlate^® 50 PM Du Pont), 0.25, 0.5, 0.75 g/l; Thiabendazole (Tecto^® 60 Syngenta) 0.25, 0.5, 0.75 g/l; and Captan (Captan^® 50 WP Adama Mexico), 1.5, 2,0, 2.5 g/l. All the fungicides used in the present research work are used in the apple growing area of the study and all of them are authorized for their use in apple fruit (^{Ramirez-Legarreta and Jacobo-Cuéllar, 2002, 172}). In vitro tests had four replications per treatment (one Petri dish was considered as one replication) and experiments, under a completely randomized design, were run three times. Control in vivo of Botrytis cinerea. Each one of twelve postharvest Golden Delicious apple fruits were wounded on the stem end of the fruit, as is a common technique used for this kind of studies (Janisiewicz et al., 2000, 1197), making a 0.5 cm diameter and 3.0 mm depth well. Wells were inoculated, before inoculation with Botrytis cinerea, with 20 µl of a 1.1x10⁹ cfu/mL suspension of each one of the three Candida oleophila strains, depending on treatment, and then 20 µl of each fungicide doses were applied in the well, of those used, according with treatment. After 15 min of the two inoculations, each well was inoculated again with 20 µl of a 1x10⁶ conidia/ml Botrytis cinerea suspension. Plastic bagged fruit was cold stored during 75 days at 0.0 °C, since is the temperature at which apple fruit is cold stored and can be stored for up to 12 months (^{Xiao and Kim, 2008, 1}). After this, the fruit was revised daily, and when damage was detected, the diameter of the lesion caused by the pathogen, in each well, was measured in millimeters. Degree of control was reported as % damage and reduction of damage due to the treatments. Experiments were run three times. The experimental design was a completely randomized one. Statistical analysis. Results from both, in vitro and in vivo treatments, were analyzed by ANOVA, under a completely randomized design using SAS (^{Statistical Analysis System, Version 6.12. Cary, NC USA}). Tukey test (α =0.05) was used when statistical differences among treatments were detected (^{Guerrero-Prieto et al., 2017, 76}).

Results

When treated with the Benomyl, Thiabendazole and Cyprodinil+Fludioxonil, the Candida oleophila strains grew with different amounts of reduction in colony growth. The Candida oleophila strains did not grow, at all, at any dose when treated with Captan. Benomyl, at all doses, did not reduce the colony growth of Candida oleophila strain L06. However, Thiabendazole and Cyprodinil+Fludioxonil did reduce the colony growth at all three doses used with statistically significant differences (α = 0.05) among them. No dose of Benomyl reduced the colony growth of L07 smooth Candida oleophila strain, but all doses of Thiabendazole and Cyprodinil+Fludioxonil resulted in a reduction with statistically significant differences (α = 0.05) among them. The colony growth for Candida oleophila strain L07 rugose was reduced by Cyprodinil+Fludioxonil at the three doses used showing statistically significant differences. Only the medium dose of Benomyl reduced the colony growth of L07 rugose. A high dose of Thiabendazole clearly resulted in the colony growth reduction of L07 rugose. Benomyl reduced the colony growth of the three Candida oleophila strains to a lesser degree (0.33 mm of damaged area when used alone, against 0.22 mm of damaged area when used combined with the Co strains), followed by Thiabendazole that had a higher rate of reduction than Benomyl and then Cyprodinil+Fludioxonil that reduced growth to a greater degree than the previous two fungicides. When the three strains of Candida oleophila were treated with Captan, none of the colonies grew (Table 1).

Candida oleophila (Co) strains in vitro colony growth reduction, was reduced, at different degree of reduction, thiabendazole, 5.3%; Benomyl, 0.3%; Cyprodinil+Fludioxonil, 28.7% and Captan, 100%, because of the fungicides and doses evaluated in the present research work. Captan did not allow any growth of any of the Candida oleophila strains at all. The mode of action of Captan, which is based on the chemical reaction between the active ingredient and sulphydric enzymes, which produces a toxic compound, thiophosgene, that interferes with the respiration of the fungi cells (^{Rusell, 2005, 16}; ^{Lima et al., 2006, 301}; ^{National Pesticide Information Center, 2018}; ^{Lima et al., 2011, 164}), resulted in a 100% reduction of the growth of the Co strains evaluated. The different amounts of colony growth reduction because of the treatments used, are also due to the mode of action of the fungicides evaluated. The fungicides evaluated in the present work, Thiabendazole, Benomyl, Cyprodinil+Fludioxonil and Captan, each one represents one of the four chemical groups of the fungicides used commercially to control fruit pathogens (^{Rusell, 2005, 16}).

In vivo Botrytis cinerea control with Candida oleophila strains and/or fungicide. Table 2 shows the results for the use of Cyprodinil+Fludioxonil and the three Candida oleophila strains. Cyprodinil+Fludioxonil and the three Candida oleophila strains resulted in a 100% control of Botrytis cinerea. When combining the strains and the fungicide, the control of Botrytis cinerea was also 100%, which indicates that the colony growth reduction shown for the Candida oleophila strains due to the combination with Cyprodinil+Fludioxonil (Table 1) did not have a reduction effect on the Botrytis cinerea damage control after treatment with Cyprodinil+Fludioxonil. Cyprodinil+Fludioxonil (Switch^® 62.5 WG Syngenta) includes two different fungicides, with different mode of action, which makes it more efficient to control, in this case, Botrytis cinerea. Cyprodinil inhibits methionine synthesis and the production of hydrolytic enzymes, which inhibits spore germination, spore germination tubes and fungi mycelia growth. Fludioxonil works on the contact surface with a preventive activity for a long time, also works on spore germination, inhibiting the penetration process and mycelia growth of the fungi and has a limited effect on fungi sporulation (^{National Pesticide Information Center, 2018}).

Table 3 includes the results when apple fruit was treated with Benomyl. Benomyl did not control Botrytis cinerea to 100%. The statistically significant (α = 0.05) reduction of Botrytis cinerea damage was 92% at a dose of 0.25 g/L and 96% when combined with L07 rugose at the same dose. At doses of 0.50 g/L, Benomyl reduced Botrytis cinerea damage by 93%. When combined with L07 smooth, the degree of reduction of damage was 98.3%. Both treatments were statistically significantly different than the Botrytis cinerea control. When the doses for Benomyl were 0.75 g/L, the greatest amount of reduction of damage for Botrytis cinerea was with Benomyl alone that resulted in an 89.5% reduction. Control of Bc damage was increased when Benomyl, from 11.8% to 8.1%, Table 3, was combined with all three Candida oleophila strains, and the degree of reduction was 94% when combined with L07 smooth, 92% when combined with L06 and 91.4% when used alone with L07 rugose. All these values were statistically equal. All three Candida oleophila strains resulted in 100% control. At doses of 1.0 g/L, Benomyl alone reduced Botrytis cinerea damage by 95.2%. However, when combined with L06 and L07 smooth, the control was 100%. An 84.7% control was obtained when combined with L07 rugose. These values were statistically equal. Treatment with Benomyl at the highest doses of 1.5 g/L, alone and combined with L06 and L07 smooth, gave 100% control. An exception occurred when Benomyl was used in combination with L07 rugose. In this case, the degree of control was 83.4%. The response of the control of the damage when Botrytis cinerea was treated with Benomyl may be due to a degree of synergism between the fungicide and the Candida oleophila strains, since the % of control was increased with an average of 67% in the different treatments used when combining Benomyl-Candida oleophila strain. The possible synergism between the Co strains and Benomyl, increased the percentage of Bc control on apple fruit, because of the modes of action on both, Benomyl, which works through carbendazim, who attaches to microtubules, interfering with several cell functions, such cell division and intracellular transport, (^{National Pesticide Information Center, 2018}) and Co strain, which use nutrient competition to control Bc (^{Guerrero-Prieto et al., 2011, 96}) and exo-ß,1-3 glucanase production (^{Guerrero-Prieto et al., 2014, 431}).

Table 4 shows the results of the Bc control when Thiabendazole was used. Thiabendazole results were like those of Benomyl since Thiabendazole alone was less effective when controlling damage of Bc, as compared when combined with the Candida oleophila strains. There was an 88.8% reduction of damage when the dose of Thiabendazole was 0.25 g/L. The degree of control increased to 100% when Thiabendazole was applied with L06 and L07 rugose and up to 91.8% when combined with L07 smooth. These values were statistically equal. Thiabendazole alone at a dose of 0.50 g/L and with all combinations, had a 100% control of the damage. A 100% of control was obtained when Thiabendazole was applied at a dose of 0.75 g/L. This was the case for both, Thiabendazole alone, and when combined with L06. However, the degree of control was 96.4% with L07 smooth and 96.8% with L07 rugose, a high degree of control, though. Treatment of Thiabendazole alone at 1.0 g/L resulted in 92% control, and a 100% control was achieved when the treatments were combined. Finally, the degree of control was 100% with Thiabendazole at doses of 1.5 g/L when L06 or L07 rugose were combined with it. Treatment with Thiabendazole alone resulted in a 90.3% control, while the degree of control was 95.2% when combined with L07 smooth. Treatments with all three Candida oleophila strains resulted in 100% control. The response of Thiabendazole was like that of Benomyl since the efficiency of control was greater when the fungicide was combined with the Candida oleophila strains. This suggests that there may be some degree of synergism between fungicide and strains, because of the modes of action on both, Benomyl, that inhibits mitosis, attaching to tubulin, avoiding fungi cell division (^{National Pesticide Information Center, 2018}) and Co strain, which use nutrient competition to control Bc (^{Guerrero-Prieto et al., 2011, 96}) and exo-ß,1-3 glucanase production, which degrades fungi cell wall and reduces spore and mycelium growth (^{Guerrero-Prieto et al., 2014, 430, 431}).

Results for the control of damage due to Botrytis cinerea when using Captan, Table 5, either alone or when combined with the Candida oleophila strains, indicate that the results observed with some of the Captan-Candida oleophila treatments in this experiment may be due solely to the use of Captan (^{Rusell, 2005, 16}, ^{Lima et al., 2006, 301}, ^{Lima et al., 2011, 164}). All three Candida oleophila strains when applied alone gave a 100% control. Captan was not compatible with the three Co strains, since did not allow any Co colony growth, as shown in Table 1. The mode of action of Captan, did not allow any colony growth of the Co strains, being toxic for the yeasts. ^{Droby et al., 2009, 138}, mention some of the key characteristics of a biocontrol agent, like to be resistant to chemicals used in the postharvest environment, however, the three Candida oleophila strains used in the present research work, were not resistant to Captan, but were resistant to the other three synthetic fungicides evaluated.

Discussion and conclusion

The use of the Candida oleophila L06, L07 smooth and L07 rugose strains as biocontrol agents on postharvest fruit, specifically in Golden Delicious apple fruit, can be considered as safe for humans who may eat the apple fruit treated with Candida oleophila (^{Droby et al., 2009, 138}). This is due to the fact of the origin of the yeasts, which are originally obtained from fresh apple fruit, since yeast belong to the natural epiphytic microflora of fruit and vegetables (^{Guerrero-Prieto et al., 2004, 224}). When yeast is used as biocontrol agents, before being applied to the produce, yeast is obtained and isolated from the fruit, then they are identified, reproduced and applied in a greater number of colonies forming units than they were originally present in the fruit surface (^{Chand-Goyal and Spotts, 1996, 254}, ^{Guerrero-Prieto et al., 2004, 224}). These yeasts can be applied to the fresh apple fruit, either by spraying them on the fruit, when still in the orchard or by immersion in the water right before cold storage and/or before fruit sorting and packing (^{Lahlali et al., 2009, 39}). This same author (^{Lahlali et al., 2009, 42}), had better control results when Pichia anomala was sprayed early in the season when the fruit was still in the apple trees in the orchard, before harvest to control Penicillium expansum.

The Candida oleophila L06, L07 smooth and L07 rugose strains can grow and can be combined with Cyprodinil+Fludioxonil, Benomyl and Thiabendazole. Treatment with the Candida oleophila strains alone and Cyprodinil+Fludioxonil alone resulted in 100% control of Botrytis cinerea. Candida oleophila L06, L07 smooth and L07 rugose, when combined with Benomyl, Thiabendazole and Cyprodinil+Fludioxonil may have some degree of synergism between each one of them to control the damage due to Botrytis cinerea on postharvest Golden Delicious apple fruit, which gives them an advantage of using a lower amount of fungicide. Candida oleophila strains could substitute the use of synthetic fungicides to control Botrytis cinerea in Golden Delicious postharvest apple fruit, thus reducing the use of synthetic fungicides, which implies a reduction on human health risk and the environment because of the use of fungicides. The contribution of the present research work is that the region's apple growers can control Botrytis cinerea damage on postharvest Golden Delicious apple fruit using Candida oleophila instead of synthetic fungicides (^{Filonow et al., 1996, 212}; ^{Santos et al., 2004, 332}; ^{Guerrero-Prieto et al., 2014, 427-428, 431}).