Biocontrol of Rice Seedling Rot Disease Caused by Curvularia lunata and Helminthosporium oryzae by Epiphytic Yeasts from Plant Leaves

Seedling rot disease in rice leads to significant loss in the production of seedlings. This research was conducted to explore yeasts that could be used as biological control agents against rice seedling rot disease caused by Curvularia lunata and Helminthosporium oryzae. In total, 167 epiphytic yeast strains were evaluated, revealing that 13 of these yeast strains demonstrated antagonistic activities against fungal pathogens and either C. lunata DOAC 2313 or H. oryzae DOAC 2293. The volatile organic compounds (VOCs) and biofilm produced were possible antagonistic mechanisms in vitro for all the antagonistic yeast strains. Using nursery trays in a greenhouse, this study evaluated the control of rice seedling rot disease caused by these two fungal pathogens using antagonistic yeasts, identified in the present study and from our previous study. Torulaspora indica DMKU-RP31 and Wickerhamomyces anomalus YE-42 were found to completely control rice seedling rot disease caused by both of these fungal pathogens. Furthermore, W. anomalus DMKU-RP04 revealed 100% disease control when the disease was caused by H. oryzae. This is the first report on using antagonistic yeasts to control rice seedling rot disease caused by C. lunata and H. oryzae. These three antagonistic yeasts also showed promising potential for development as biocontrol agents against rice seedling rot disease caused by fungi.


Introduction
Rice (Oryza sativar) is the most widely produced and consumed staple food in Asian countries [1]. In 2018, the Rice Department in Thailand's Ministry of Agriculture and Cooperatives reported that Thailand had approximately 59.2 million hectares used for rice cultivation with rice production of 24.2 million tons [2].
In rice cultivation, seeds are germinated in nursery trays and seedlings are transplanted to fields. During rice seed germination and the growth of seedlings, various fungal pathogens may cause diseases in seeds and seedlings [3]. One of the significant diseases in rice seeds and seedlings is seedling rot, caused by various seed-borne and soil-borne fungal pathogens, such as Curvularia lunata, Fusarium oxysporum, Helminthosporium oryzae, Pyricularia oryzae, Pythium spp., and Rhizoctonia solani [2,[4][5][6][7]. In Thailand, two of the significant causal agents of rice seedling rot in nursery trays are C. lunata and H. oryzae, as reported by the Rice Department, Ministry of Agriculture and Cooperatives of Thailand [2]. Rice seedling rot disease affects seeds and seedlings of rice plants, with disease symptoms often appearing within a few days of sowing the seeds. Infected seeds become soft and pulpy, and may be surrounded by white mold growth, while some cannot germinate. From the infected seeds that manage

Evaluation In Vitro of Antagonistic Activities of Epiphytic Yeasts against Fungal Pathogens Causing Rice Seedling Rot Disease
Using dual cultivation [21] with slight modification, the 167 epiphytic yeast strains were evaluated for their antagonistic activities against C. lunata DOAC 2313 and H. oryzae DOAC 2293. For each strain, an active yeast culture obtained by cultivation on YM agar at 25 • C for two days was inoculated on potato dextrose agar (PDA) [Difco™-BBL, Sparks, MD, USA] in a Petri dish by linear streaking 3 cm from the dish edge and incubated at 25 • C for two days. A fungal mycelial plug (5 mm in diameter), obtained from an active fungal pathogen colony seven days at 25 • C on PDA agar) and cut by a cork borer, was placed on the opposite edge of the Petri dish to that previously inoculated with yeast. A PDA dish inoculated only with the fungal pathogen was used as a control. The inoculated dishes were incubated at 25 • C for seven days. Three replicates were performed for each treatment, and the experiment was repeated three times. The percentage of fungal pathogen growth inhibition was calculated as follows: Growth inhibition (%) = (radius of fungal pathogen colony cultured alone-radius of fungal pathogen colony culture with yeast)/radius of fungal pathogen colony cultured alone × 100.

Production of Antifungal Volatile Organic Compounds (VOCs)
The volatile organic compound (VOC) production of the antagonistic yeasts was determined by a double dish assay [22] with slight modification, as described by Into et al. [13]. In brief, an aliquot (100 µL) of yeast cell suspension (10 8 colony forming units [CFU]/mL) was spread on a PDA dish and incubated at 25 • C for two days, while another PDA dish, inoculated with a fungal mycelial plug, was placed upside down instead of the dish cover. The control was a PDA dish inoculated only with fungal pathogens. Three replicates were performed for each treatment and control. The percentage of fungal pathogen growth inhibition was calculated as follows: Growth inhibition (%) (diameter of fungal pathogen colony cultured alone-diameter of fungal pathogen colony culture with yeast)/diameter of fungal pathogen colony cultured alone × 100.

Competition for Nutrients
Using the dual cultivation method described by Zhang et al. [23], the competition for nutrients was estimated by the amount of fungal pathogen growth inhibition by each antagonistic yeast in a PDA medium containing different nutrient concentrations. The evaluation was performed in the same way as in Section 2.3, but PDA media with four different nutrient concentrations were used. These consisted of: standard nutrient concentration; half of standard nutrient concentration; one-quarter of standard nutrient concentration; and one-tenth of standard nutrient concentration prepared from 39 g/L, 19.5 g/L, 9.7 g/L, and 3.9 g/L Difco™-BBL PDA powder, respectively. Three replicates for each treatment and control were carried out.

Production of β-Glucanase and Chitinase
Production of β-glucanase and chitinase by the antagonistic yeasts was carried out by their cultivation in potato dextrose broth (PDB) [Difco™-BBL, Sparks, MD, USA] at 150 rpm and 25 • C for five days, as described by Into et al. [13]. The culture broth was collected and centrifuged at 10,000× g for 5 min. The supernatant was analyzed for β-glucanase and chitinase activities.
The activities of β-glucanase and chitinase were determined by the colorimetric quantification of reducing sugar and N-acetyl glucosamine (NAG) released from laminarin and colloidal chitin, respectively, as described by Into et al. [13]. The concentrations of reducing sugar and N-acetyl glucosamine were determined by Miller's [24] method, with β-glucanase and chitinase activities expressed as units (U) per mL. One unit (U) of β-glucanase was defined as 1 µg of reducing sugar released from laminarin per minute under the assay conditions, whereas one unit (U) of chitinase was defined as 1 µg of NAG released from colloidal chitin per minute also under the assay conditions.

Biofilm Formation
The biofilm formation of the antagonistic yeast strains was investigated using Růžička et al.'s [25] method with slight modification, as described by Into et al. [13]. In brief, an aliquot (20 µL) of yeast cell suspension (cells grown on PDA at 25 • C for two days suspended in sterile water and adjusted to an optical density (OD) measured at 600 nm of 0.5) was inoculated into each well of a 96-well microtiter plate containing 180 µL of PDB. The microtiter plate was incubated at 25 • C for 48 h (h). A well containing only PDB was used as a negative control. Three replicates were performed for each treatment. After 48 h, the wells were emptied, rinsed with water, and air-dried at room temperature.
The adherent biofilm layer was stained with an aqueous solution of 1% (w/v) crystal violet for 20 min, rinsed with water, and air-dried. The stained biofilm layer was eluted from each well with 200 µL of ethanol, and the OD of each well was measured at 620 nm. Biofilm formation in a well was regarded as positive when the mean OD of the treatment was higher than the mean OD of the negative control (ODc). The following classification was applied in the determination of biofilm formation: weak biofilm producer (ODc < ODt ≤ 2 ODc); moderate biofilm producer (2 ODc < ODt ≤ 4 ODc); and strong biofilm producer (4 ODc < ODt) [26].

Siderophore Production
Siderophore production by the antagonistic yeasts was investigated by their cultivation on chrome azurol S (CAS) [Sigma-Aldrich Chemie GmbH, Schnelldorf, Germany] blue agar in a Petri dish [27]. An aliquot of yeast cell suspension was prepared by suspending a loopful of culture grown on YM agar at 25 • C for two days in 3 mL of 0.85% sterile normal saline, adjusted to an OD 600 of 0.10. An aliquot (10 µL) of yeast cell suspension was dropped onto the surface of CAS blue agar and incubated in the dark at 25 • C for 10 days. Three replicates were performed for each treatment. Siderophore production was confirmed by a change in the color of the medium from blue to purple or yellow around the yeast colony.

Phosphate and Zinc Oxide Solubilization
Phosphate and zinc oxide solubilization by the antagonistic yeasts was determined by their cultivation on Pikovskaya's agar [28] and zinc oxide agar [29], respectively. An aliquot (5 µL) of yeast cell suspension (OD 600 of 0.10), prepared in the same way as in Section 2.4.5, was dropped onto the surface of Pikovskaya's agar or zinc oxide agar in a Petri dish and incubated at 25 • C for five days. Three replicates were performed for each treatment. The solubilization efficiency (SE) of the phosphate and zinc oxide was designed as a ratio of the halo zone diameter and the colony diameter, with both these diameters measured.
In the present study, the Thai Jasmine rice cultivar, Khao Dawk Mali 105, was used. Rice seeds (15 g) were surface sterilized by 100 mL of 10% Clorox solution for one minute, rinsed with sterile distilled water, and air dried at room temperature.
Each antagonistic yeast was cultivated in yeast extract peptone dextrose (YPD) broth (comprised of 20 g/L glucose, 20 g/L peptone, and 10 g/L yeast extract) on a rotary shaker at 150 rpm, at 25 • C for 48 h (h). Yeast cell suspension was prepared by suspending cells collected from the culture broth by centrifugation at 5000× g for 10 min in the sterile Ringer solution. Cells were quantified with a hemacytometer to reach a concentration of 10 8 cells/mL. Fungal pathogens were grown on PDA at 25 • C for seven days, with the spore suspension (10 5 spores/mL) prepared in sterile 0.05% Tween 20. Soil was put in a plastic bag and sterilized in an autoclave at 121 • C for 30 min. The sterilization was performed twice.
Infected rice seed was prepared, as described by Khalili et al. [12]. Sterilized rice seeds were soaked in the spore suspension (10 5 spores/mL) of each fungal pathogen and then shaken on a rotary shaker (TAITAC BR-300LF, Saitama, Japan), at 100 rpm, at room temperature (28-32 • C) for 24 h and then air-dried. The infected seeds were soaked in the yeast cell suspension (10 8 cells/mL) for two hours; 100 treated seeds were then transplanted into a plastic tray (20 cm wide × 35 cm long × 10 cm high) containing sterile soil (4 kg). The seeds were sown in four rows: in each row, 16-17 seeds were placed, with each seed 2-3 cm apart from the other seeds. The tray was incubated in greenhouse conditions (11 h light/13 h dark, and at 30-34 • C day/20-24 • C night) for 45 days. Three replicates were performed for each treatment (3 × 100 seeds). The number of germinated seeds, and seedlings' stem height, root length and dry weight were recorded.

Statistical Analysis
Data were analyzed with statistical analysis software IBM ® SPSS Statistics version 22 (Armonk, NY, USA). All data were first subjected to analysis of variance (ANOVA). Means were compared using Duncan's multiple range test and a significance level of p ≤ 0.05 was considered as being significantly different.

Production of Antifungal Volatile Organic Compounds (VOCs)
The present study evaluated the antagonistic activities against C. lunata DOAC 2313 and H. oryzae DOAC 2293 by VOCs production form 11 and six antagonistic yeast strains, respectively. The results revealed that all antagonistic yeast strains produced VOCs inhibited growth of their related fungal pathogens but with lower inhibition percentages than using yeast cultures ( Table 1). The 10 antagonistic yeast strains tested inhibited C. lunata DOAC 2313 growth by 4.2-32.1% and W. anomalus YE-42 produced VOCs that showed the strongest inhibitory effect on this fungal pathogen. On the other hand, the six antagonistic yeast strains tested with H. oryzae DOAC 2293 produced VOCs inhibited this fungal pathogen in the range of 0.7% to 25.1% and W. anomalus YE-42 revealed the strongest antifungal VOCs activity. However, H. sinensis YE-19, which showed the strongest inhibitory effect when tested by dual cultivation with H. oryzae DOAC 2293 produced VOCs inhibited this fungal pathogen only by 1%.    1 Inhibition (%) = (Diameter of fungal colony grow alone-Diameter of fungal colony grow with yeast/Diameter of colony grow alone) ×100; Each value represents a mean " ± " standard deviation (SD). In the same column for each rice pathogenic fungus tested data followed by the different lower-case letters are significantly different according to Duncan's multiple range test at p ≤ 0.05. 2 Inhibition (%) = (Radius of control fungal colony-Radius of fungal colony grow with yeast/Radius of control fungal colony) ×100; Each value represents a mean " ± " standard deviation (SD). In the same row data followed by the different lower-case letters are significantly different according to Duncan's multiple range test at p ≤ 0.05. 3 Standard nutrient concentration (39 g/L PDA powder). 4 Half of standard nutrient concentration (19.5 g/L PDA powder). 5 One-fourth of standard nutrient concentration (9.7 g/L PDA powder). 6 One-tenth of standard nutrient concentration (3.9 g/L PDA powder).

Production of β-Glucanase and Chitinase
The production of β-glucanase and chitinase by the 13 antagonistic yeast strains was determined by their cultivation in PDB at 25 • C for five days with observation of the enzymatic activities of cell-free culture broth. The results revealed that β-glucanase was produced by 11 strains: however, only a small quantity (0.2-116.5 mU/mL) was produced, with R. mucilaginosa YE-171 producing the highest amount ( Table 2). Eight antagonistic yeast strains produced small amounts (1.0-504.5 mU/mL) of chitinase, while P. japonica YE-135 produced the largest amount.

Biofilm Formation
The biofilm formation of the 13 antagonistic yeast strains was determined, with the results indicating that all 13 strains formed biofilm. Of the 13 strains, eight revealed moderate biofilm formation, while four showed weak biofilm formation. The only strain with strong biofilm formation was C. tropicalis YE-111 (Table 3).  1 The optical density at 620 nm of biofilm layer stained with crystal violet (OD T ), data express as average OD T ± SD). 2 Average optical density of biofilm layer (OD T ) as a portion of optical density of control (OD C ). 3 Interpretation of biofilm formation: OD T ≤ Ac, no biofilm formation; weak biofilm producer (OD C < OD T ≤ 2 OD C ), moderate biofilm producer (2 OD C < OD T ≤ 4 OD C ) and strong biofilm producer (4 OD C < OD T ), when OD C = 0.0580). 4 Diameter of holo zone (cm). e Solubilization efficiency (SE) = Diameter of the halo zone (cm)/Diameter of the colony (cm).

Phosphate and Zinc Oxide Solubilization
The phosphate and zinc oxide solubilizing activity of all antagonistic yeasts were determined on Pikovskaya s agar and zinc oxide agar, respectively, showing that only three strains of W. anomalus grew and produced halo zones around colonies. The phosphate and zinc oxide solubilization efficiency (SE) units were in the range of 1.3-1.5 (Table 3, Figure 4).

Evaluation of the Control of Rice Seedling Rot Disease in a Greenhouse
The efficacy of antagonistic yeasts, obtained from the present study and the previous study, in the control of rice seedling rot disease caused by C. lunata DOAC 2313 and H. oryzae DOAC 2293 were evaluated in nursery trays in a greenhouse. The two chemical fungicides, carbendazim ® and mancozeb ® , were used for comparison.
To test the control of rice seedling rot disease caused by C. lunata DOAC, 19 antagonistic yeast strains were tested. T. indica DMKU-RP31 and W. anomalus YE-42 were found to show higher seedling vigor index (SVI) percentages than the negative control and carbendazim ® : these two strains resulted in no disease incidence and, consequently, 100% disease control (Table 4, Figure 5). Fourteen of the 19 antagonistic yeast strains showed no significantly different SVI percentages when compared with the negative control and carbendazim ® : they resulted in disease incidence ranging from 2.17%-16.57% and, consequently, disease control of 45.70%-92.88%. The remaining four yeast strains revealed lower SVI percentages than the negative control and carbendazim ® : they resulted in low disease control ranging from 33.19%-42.50%. When using the two chemical fungicides, carbendazim ® and mancozeb ® , the study revealed 6.27% and 30.48% disease incidence and, consequently, 79.47% and 0.15% disease control, respectively.

Evaluation of the Control of Rice Seedling Rot Disease in a Greenhouse
The efficacy of antagonistic yeasts, obtained from the present study and the previous study, in the control of rice seedling rot disease caused by C. lunata DOAC 2313 and H. oryzae DOAC 2293 were evaluated in nursery trays in a greenhouse. The two chemical fungicides, carbendazim ® and mancozeb ® , were used for comparison.
To test the control of rice seedling rot disease caused by C. lunata DOAC, 19 antagonistic yeast strains were tested. T. indica DMKU-RP31 and W. anomalus YE-42 were found to show higher seedling vigor index (SVI) percentages than the negative control and carbendazim ® : these two strains resulted in no disease incidence and, consequently, 100% disease control (Table 4, Figure 5). Fourteen of the 19 antagonistic yeast strains showed no significantly different SVI percentages when compared with the negative control and carbendazim ® : they resulted in disease incidence ranging from 2.17%-16.57% and, consequently, disease control of 45.70%-92.88%. The remaining four yeast strains revealed lower SVI percentages than the negative control and carbendazim ® : they resulted in low disease control ranging from 33.19%-42.50%. When using the two chemical fungicides, carbendazim ® and mancozeb ® , the study revealed 6.27% and 30.48% disease incidence and, consequently, 79.47% and 0.15% disease control, respectively. Each value represents a mean " ± " standard deviation (SD). In the same column, data followed by the different lower-case letters are significantly different according to Duncan's multiple range test at p ≤ 0.05. 1 Antagonistic yeasts from previous investigation [13]. 2 Dry weight of 100 rice plants nd = not determined. Evaluation of 11 antagonistic yeast strains was conducted to investigate if they had an inhibitory effect on H. oryzae DOAC 2293 to control rice seedling rot disease which it caused. T. indica DMKU-RP31, W. anomalus DMKU-RP04, and W. anomalus YE-42 showed higher SVI percentages than the negative control and carbendazim ® ; these strains resulted in no disease incidence and, consequently, 100% disease control (Table 5, Figure 6). Six antagonistic strains revealed no SVI percentages that were significantly different and, consequently, disease control ranged from 75.27%-97.45%. The remaining three strains showed lower SVI percentages than the negative control and carbendazim ® and, consequently, 44.95%-63.37% disease control. Mancozeb ® revealed the lowest disease control (of 3.05%) whereas carbendazim ® showed a high level of disease control at 94.04%. Evaluation of 11 antagonistic yeast strains was conducted to investigate if they had an inhibitory effect on H. oryzae DOAC 2293 to control rice seedling rot disease which it caused. T. indica DMKU-RP31, W. anomalus DMKU-RP04, and W. anomalus YE-42 showed higher SVI percentages than the negative control and carbendazim ® ; these strains resulted in no disease incidence and, consequently, 100% disease control (Table 5, Figure 6). Six antagonistic strains revealed no SVI percentages that were significantly different and, consequently, disease control ranged from 75.27%-97.45%. The remaining three strains showed lower SVI percentages than the negative control and carbendazim ® and, consequently, 44.95%-63.37% disease control. Mancozeb ® revealed the lowest disease control (of 3.05%) whereas carbendazim ® showed a high level of disease control at 94.04%. Each value represents a mean " ± " standard deviation (SD). In the same column, data followed by the different lower-case letters are significantly different according to Duncan's multiple range test at p ≤ 0.05. 1 Antagonistic yeasts from previous investigation [13]. 2 Dry weight of 100 rice plants. nd = not determined.

Discussion
The results of this study revealed that the yeast species showing antagonistic activity against the two fungal pathogens, C. lunata DOAC 2313 and H. oryzae DOAC 2293, the causes of rice seedling rot disease, were from both yeast phyla, namely, phyla Ascomycota (C. michaelii, C. tropicalis, and W. anomalus) and Basidiomycota (H. sinensis, P. japonica, P. hubeiensis, R. mucilaginosa, and R. taiwanensis). Some antagonistic yeast species found in the present study have previously been reported to possess antagonistic activities. W. anomalus was reported for its ability to antagonize various fungal pathogens, the causes of plant and post-harvest diseases, such as Alternaria alternata, Aspergillus carbonarius, C. lunata, Botrytis cinerea, Fusarium moniliforme, Helmintosporium oryzae, Monilinia fructicola, Penicillium expansum, Rhizoctonia solani, Cladosporium spp., and Colletotrichum spp. [14,30,31]. C. tropicalis revealed the ability to control black rot in pineapple caused by Chalara paradoxa [32]. Cryptococcus albidus was found to be the effective antagonistic yeast against P. expansum and apple blue mold [30]. H. sinensis and R. taiwanensis were reported for their antagonistic activities against Aspergilus flavus [33,34]. R. mucilaginosa demonstrated the ability to antagonize Botrytis cinerea, the cause of gray mold spoilage in strawberries [35] and Penicillium expansum, the cause of postharvest apple disease [36]. P. hubeiensis could control Lasiodiplodia theobromae [15].
In the present study, the antagonistic mechanisms of 13 yeast strains from eight species were evaluated in vitro, with the results revealing that all strains possessed the ability to produce antifungal VOCs and biofilm to antagonize C. lunata and H. oryzae. Some strains produced both of the cell wall lytic enzymes, β-glucanase and chitinase, some strains produced only β-glucanase, whereas some did not produce either enzyme. The production of cell wall lytic enzymes appeared to

Discussion
The results of this study revealed that the yeast species showing antagonistic activity against the two fungal pathogens, C. lunata DOAC 2313 and H. oryzae DOAC 2293, the causes of rice seedling rot disease, were from both yeast phyla, namely, phyla Ascomycota (C. michaelii, C. tropicalis, and W. anomalus) and Basidiomycota (H. sinensis, P. japonica, P. hubeiensis, R. mucilaginosa, and R. taiwanensis). Some antagonistic yeast species found in the present study have previously been reported to possess antagonistic activities. W. anomalus was reported for its ability to antagonize various fungal pathogens, the causes of plant and post-harvest diseases, such as Alternaria alternata, Aspergillus carbonarius, C. lunata, Botrytis cinerea, Fusarium moniliforme, Helmintosporium oryzae, Monilinia fructicola, Penicillium expansum, Rhizoctonia solani, Cladosporium spp., and Colletotrichum spp. [14,30,31]. C. tropicalis revealed the ability to control black rot in pineapple caused by Chalara paradoxa [32]. Cryptococcus albidus was found to be the effective antagonistic yeast against P. expansum and apple blue mold [30]. H. sinensis and R. taiwanensis were reported for their antagonistic activities against Aspergilus flavus [33,34]. R. mucilaginosa demonstrated the ability to antagonize Botrytis cinerea, the cause of gray mold spoilage in strawberries [35] and Penicillium expansum, the cause of post-harvest apple disease [36]. P. hubeiensis could control Lasiodiplodia theobromae [15].
In the present study, the antagonistic mechanisms of 13 yeast strains from eight species were evaluated in vitro, with the results revealing that all strains possessed the ability to produce antifungal VOCs and biofilm to antagonize C. lunata and H. oryzae. Some strains produced both of the cell wall lytic enzymes, β-glucanase and chitinase, some strains produced only β-glucanase, whereas some did not produce either enzyme. The production of cell wall lytic enzymes appeared to be a strain-specific characteristic, not a species-specific characteristic. The antagonistic yeast strains of four species, namely, P. hubiensis, R. mucilaginosa, R. taiwanensis, and W. anomalus produced siderophores, with W. anomalus also capable of phosphate and zinc oxide solubilization. Therefore, the antagonistic mechanisms of all the yeast strains were found to emit antagonistic antifungal VOCs and to form biofilm. Some yeast strains could possibly have additional mechanisms.
As shown in the present study, the production of antifungal VOCs, which are low molecular weight chemical compounds with high vapor pressure, played a role in the antagonistic activity of C. michaelii, C. tropicalis, H. sinensis, P. japonica, P. hubeiensis, R. mucilaginosa, R. taiwanensis, and W. anomalus. This result was in agreement with other investigators' results which showed that the emission of VOCs by antagonistic yeasts has proven to be one of the important direct antagonistic mechanisms against various fungal pathogens of antagonistic yeasts [14,33,[37][38][39]. Debaryomyces nepalensis produced VOCs to control the pathogen Colletotrichum gloeosporioides in mango fruit [40]. Metschnikowia pulcherrima was found to produce VOCs that could suppress the growth of C. gloesporioides [41]. T. indica and P. hubeiensis were reported to produce VOCs that inhibited the growth of L. theobromae, whereas P. aspenensis produced VOCs that could inhibit the growth of C. gloeosporioides [15]. W. anomalus, M. pulcherrima, and S. cerevisiae produced VOCs capable of decreasing the mycelial growth of B. cinerea, M. fructicola, A. alternata, A. carbonarius, P. digitatum, Cladosporium spp., and Colletotrichum spp. [31]. W. anomalus produced VOCs which inhibited the growth of C. lunata, F. moniliforme, and R. solani, whereas K. ohmeri produced VOCs which inhibited only the growth of R. solani [14]. T. indica produced VOCs to strongly inhibit the growth of R. solani and Pyricularia oryza [13].
In this study, no antagonistic yeasts had the characteristic of competition for nutrients when dual cultured with C. lunata DOAC 2313 and H. oryzae DOAC 2293. However, some yeast species were reported to compete for nutrients with the pathogens. For example, with P. hubeiensis and T. indica, it was revealed that competition for nutrients resulted in inhibition in the growth of L. theobromae [40]. M. pulcherrima was reported to control C. gloeosporioides through its ability to compete for nutrients [41].
All the antagonistic yeast strains tested could form biofilm to different degrees. This means that competition for space was one of the antagonistic mechanisms of these antagonistic yeasts. Competition for space is based on biofilm formation. Some yeasts form biofilm, allowing them to adhere to surfaces, colonize, and resist stresses [42]. W. anomalus and P. hubeiensis were previously reported to form biofilm [14,15,43]. In addition, some other yeast species were reported to form biofilm that played a role in biological control activities, such as K. ohmeri, Pichia fermentans, P. aspenensis, and T. indica [15,44].
The results of the present study revealed that β-glucanase and chitinase were produced by some strains of C. michaelii, C. tropicalis, H. sinensis, P. japonica, P. hubiensis, R. mucilaginosa, R. taiwanensis, and W. anomalus. This indicated that the fungal cell wall lytic enzymes of some strains of these yeast species could be one of the mechanisms for controlling rice seedling rot fungal pathogens. Some antagonistic yeasts were reported to produce cell wall lytic enzymes which could play a role in antagonistic activity, for example, W. anomalus, K. ohmeri, M. caribbica, Meyerozyma guilliermondii, M. pulcherrima, and T. indica [13,[45][46][47].
Siderophores are involved in competition for the nutrient iron (Fe +3 ) [48]. When antagonistic microorganisms have this ability, they can use the iron; therefore, the iron in the substrate is depleted, resulting in limited growth of the pathogens. The result of this study showed that seven antagonistic yeast strains in four species, P. hubiensis, R. mucilaginosa, R. taiwanensis, and W. anomalus, produced siderophores. Therefore, producing siderophores could be one of the antagonistic mechanisms of these antagonistic yeasts. P. hubeiensis and W. anomalus have previously been reported to produce siderophores [13][14][15]. In addition, other yeast species were found to produce siderophores, such as P. aspenensis, Rhodotorula glutinis, and T. indica [15,49].
Solubilization of phosphate and zinc oxide can convert them to a soluble form, easily assimilated by plants, that promotes plant growth [50,51]. The results of the present study indicated that phosphate and zinc oxide solubilization was among the indirect mechanisms of W. anomalus, with strains of this species having previously been reported to have the ability to solubilize phosphate and zinc oxide [13,14].
The evaluation of the control of rice seedling rot disease caused by C. lunata DOAC 2313 and H. oryzae DOAC 2293 revealed that all the antagonistic yeast strains tested could control this disease. Of these antagonistic yeasts, T. indica DMKU-RP31, obtained from Into et al.'s [13] previous study, and W. anomalus YE-42, derived in the present study, could completely control the disease caused by both fungal pathogens. On the other hand, W. anomalus DMKU-RP04, obtained from the previous study, could completely control the disease only when infected with C. lunata DOAC 2313. These three yeast species demonstrated better control of rice seedling rot disease than the effective chemical fungicide, carbendazim ® . Interestingly, T. indica DMKU-RP31 was recently reported to control fruit disease in post-harvest ripe mangos caused by Lasiodiplodia theobromae [15].
Reports on the management of diseases in rice seed and seedlings to date are limited. Chemical fungicides have been used to control rice seed and seedling diseases caused by different fungal pathogens, with these including carbendazim ® , difolatan ® , mancozeb ® , propiconazole ® , and validamycin ® [52][53][54][55][56]. In the present study, two chemical fungicides, carbendazim ® and mancozeb ® , were used for comparison with antagonistic yeasts in the control of rice seedling rot disease caused by C. lunata and H. oryzae. Mancozeb ® was found to have very weak controlling ability, 0.15% and 3.05% disease control, for C. lunata and H. oryzae, respectively. The reason may be that mancozeb ® inhibited rice seed germination [57]. The management of rice seedling rot disease by using soil amended with tricin-releasing rice hulls was reported, with this showing suppression of the soil-borne fungal pathogens viz. F. oxysporum and R. solani that caused rice seedling rot disease [Kong et al., 2010]. The reason was that ricin was reported as being detected in rice hulls and that it had fungicidal activity. Few reports were found on using antagonistic microorganisms to control rice seedling rot disease. Research has explored the use of antagonistic bacteria, Pseudomonas fluorescens and Pseudomonas tolaasii, to control rice seedling rot disease caused by Achlya klebsiana and Pythium spinosum [11]. However, for yeasts, the only report found used Metschnikowia pulcherrima and Pichia guilliermondii to control Fusarium fujikuroi caused by bakanae disease in rice seeds [20]. To the best of our knowledge, the report on the present study is the first one on the use of yeasts for the control of rice seedling rot disease caused by C. lunata and H. oryzae.

Conflicts of Interest:
The authors declare that there are no conflict of interest.