Evaluation of yeasts obtained from Antarctic soil samples as biocontrol agents for the management of postharvest diseases of apple (Malus × domestica)

Authors


Correspondence: Silvana Vero, Cátedra de Microbiología, Departamento de Biociencias, Facultad de Química, UdelaR, Gral. Flores 2124, Montevideo, CC 11800, Uruguay. Tel.: +598 292 44209; fax: +598 292 41906; e-mail: svero@fq.edu.uy

Abstract

Psychrotrophic yeasts were isolated from Antarctic soils, selected based on their ability to grow in apple juice at low temperatures, and were evaluated as potential biocontrol agents for the management of postharvest diseases of apple during cold storage. Among the species recovered, an isolate of Leucosporidium scottii, designated At17, was identified as a good biocontrol agent for blue and gray mold of two apple cultivars. The selected isolate produced soluble and volatile antifungal substances that were inhibitory to apple pathogens. Siderophore production was also demonstrated, but it did not appear to play a role in pathogen inhibition. The selected yeast had the capacity to form a biofilm when grown in apple juice, which is considered an important attribute of postharvest antagonists to successfully colonize wounds and intact fruit surfaces. At17 was resistant to commonly used postharvest fungicides, so application of a combination of low-dose fungicide along with the biocontrol agent could be used as an integrated management practice.

Introduction

After harvest, apples are stored at low temperature (0–1 °C) to maintain quality and to minimize spoilage. However, development of fungal diseases, caused mainly by Pencillium expansum and Botrytis cinerea, cannot be avoided. Currently, the control of postharvest rots has relied mainly on the use of synthetic fungicides. As fungicidal resistance has become more prevalent and consumer demands for the reduction in chemical residues on fruit are increasing, alternative disease control strategies are needed. In this regard, biological control has developed as a potential alternative. In particular, the use of yeast as biocontrol agents of postharvest pathogens has been the subject of a great deal of research (Droby et al., 2009). In the case of apples, various yeast species have been demonstrated to be well adapted to the immediate environment and nutritional conditions prevailing at the wound site and therefore able to competitively colonize apple wounds, preventing pathogen establishment and development (Pimenta et al., 2009; Nally et al., 2012).

Many mechanisms acting in concert have been demonstrated to be responsible for the biocontrol activity of yeasts (Droby et al., 2009). Wound colonization and nutrient competition appear as the primary mechanisms; however, inhibition of pathogen development by soluble or volatile metabolites or by direct hyphal parasitism and induction of host resistance have also been reported (Wisniewski et al., 1991; Droby et al., 2002; Mohamed & Saad, 2009; Tongsri & Sangchote, 2009; Huang et al., 2011). Additional attributes of yeasts have been associated with their biocontrol activity. For example, the ability to form biofilms is clearly involved in adhesion to specific surfaces, colonization, and resistance to different stresses (Harding et al., 2011). Additionally, the ability to tolerate high levels of reactive oxygen species produced by fruit tissue is an essential characteristic of effective yeast antagonists (Castoria et al., 2003; Macarisin et al., 2010). In the case of apple, the ability of yeasts to grow at the low pH and high sugar concentrations found in fruit wound sites is also a characteristic necessary to ensure wound colonization. Furthermore, the production of specific metabolites, such as siderophores (i.e. iron-chelating compounds), may be associated with nutrient compe-tition (Calvente et al., 2001), and the excretion of lytic enzymes (chitinases, glucanases, proteases) may be involved in direct parasitism or inhibition of pathogen development (Droby et al., 2009).

As harvested apples are placed in cold storage for varying amounts of time, an important attribute to consider in the selection of yeasts as biocontrol agents of apple postharvest diseases is the ability to be effective at low temperatures. In this regard, the use of psychrotrophic yeast may be a distinct advantage. In previous research (Vero et al., 2009, 2011), biocontrol agents were isolated from the superficial microflora present on fruit stored for at least 3 months at low temperature. Washes from these fruit were plated on petri plates containing either apple juice agar or potato dextrose agar (PDA) and incubated for 10 days at 5 °C to favor the isolation of psychrotrophic microorganisms. Microbial diversity was low, which simplified the process of identification and evaluation. Only a few species were recovered, but all were well adapted to cold storage conditions. Most isolates exhibited some degree of biocontrol activity against apple postharvest diseases when tested on fruit.

In this report, we describe our efforts to develop an approach to efficiently identify yeast species that are both well adapted to low temperatures and also have the potential to serve as postharvest biocontrol agents. Our approach entailed the evaluation of psychrotrophic yeasts isolated from soils collected in Antarctica based on the ability to grow in apple juice at low temperatures. It was felt that this technique would enable us, in a more direct and less laborious manner, to identify potential postharvest biocontrol agents that could be used to treat apples in cold storage. Among the species recovered, the At17 isolate of Leucosporidium scottii was identified as a good biocontrol agent for blue and gray mold of two apple cultivars. The biocontrol activity of At17 was compared with a previously selected antagonist (PL1) isolated from fruit (Vero et al., 2009, 2011). Additionally, potential mechanisms associated with biocontrol activity were studied, and the compatibility of At17 with fungicides or diphenylamine was evaluated.

Materials and methods

Isolation and identification of potential biocontrol agents

Microbial isolates were obtained from 21 soil samples collected in the surroundings of the Uruguayan scientific research station in Antarctica (General Artigas Base) located on King George Island (Lat: 62°11′04 South Long: 58°54″ West), about 100 km from the Antarctic Peninsula. The temperature of the soil samples was between 0 and 2 °C at the time of collection. Samples were kept at 1 °C until they were sampled for microflora. One gram of soil from each sample was placed in 50-mL Erlenmeyer flasks containing 10 mL of sterile apple juice and incubated at 9 ± 1 °C for 10 days in orbital shaker at 150 r.p.m. After incubation, a small sample of liquid was removed using a microbial sample loop from cultures that showed signs of microbial growth as evidenced by turbidity. Individual samples were streaked on the surface of apple juice agar plates. Plates were then incubated at 9 ± 1 °C for 7–10 days. Colonies were picked from these plates and streaked on to PDA plates to obtain pure cultures.

All isolates were identified to species level by sequencing the ITS1-ITS2 and the D1/D2 genomic regions in both directions. DNA extraction and amplification of the ITS1-ITS2 region were carried out according to Vero et al. (2009). D1/D2 amplification was performed as described in Kurtzman & Droby (2002). Sequencing of the purified PCR products was performed at Macrogen Inc. (Seoul, Korea). Sequence similarity searches were performed with blast network service of the NCBI database (http://blast.ncbi.nlm.nih.gov/Blast.cgi).

Pathogens

Native strains of P. expansum and B. cinerea, selected for their high level of aggressiveness on Red Delicious apples (Vero et al., 2009), were used in this study. Cultures were maintained on PDA at 5 °C.

Selection of biocontrol agents

Selection of biocontrol agents was based on their ability to prevent blue and gray mold, caused by P. expansum and B. cinerea, respectively, in inoculated wounds of ‘Red Delicious’ and ‘Pink Lady’ apples, at 1.0 ± 0.5 °C. Biocontrol activity of isolates was compared with that of PL1, a yeast antagonist identified as Cystofilobasidium infirmominiatum, previously isolated from fruit in cold storage and demonstrated to control postharvest decay in apples (Vero et al., 2011). Within 1 week of harvest, apples were surface-disinfected with sodium hypochlorite (0.1%) for 2 min and then rinsed in running tap water. Five wounds (3 mm deep × 2 mm wide) were made along the equator of each apple. Wounds were then inoculated with 10 μL of a microbial suspension (107 CFU mL−1) of the potential antagonist or 10 μL of sterile water in control apples. Microbial suspensions were prepared in 5 mL of sterile water, using 2-day-old PDA cultures to reach a concentration of 107 CFU mL−1. After 2 h, wounds were inoculated with 10 μL of a conidial suspension (104 conidia mL−1) of either pathogen. This pathogen concentration had been previously shown to produce 100% infection of wounds (Vero et al., 2002). Five apples per treatment were used, and treatments were replicated three times in different trays. Apples were stored in a cold chamber at 1.0 ± 0.5 °C for 3 months after which disease incidence (measured as the percentage of wounds infected) was evaluated. Experiments were repeated twice in two different years.

The Antarctic yeast isolate that exhibited the highest level of disease control against both pathogens in both apple cultivars was selected for further study. An additional experiment was conducted to determine the ability of a 50/50 (v/v) mixture of isolate PL1 and the selected Antarctic isolate to control P. expansum on ‘Red Delicious’ apples. Yeast suspensions of each isolate were prepared as described to obtain a concentration of 1 × 107 CFU mL−1. To prepare the 50/50 mixture, an equal volume of the two yeast suspensions was mixed. A control assay using each antagonist alone was performed in parallel.

Phylogeny and phenotypic characterization of the selected biocontrol agent

Phylogenetic analyses of D1/D2 and ITS1-ITS2 sequences of the selected isolate were conducted using mega version 5 (Tamura et al., 2011). DNA sequences were aligned with sequences of homologous regions of closely related species (Leucosporidium spp., Leucosporidiella spp., and Mastigobasidium intermedium) retrieved from GenBank. Rhodotorula glutinis and Rhodotorula graminis were designated as outgroup species in both analyses.

Evolutionary distances were computed using the Jukes–Cantor method (Jukes & Cantor, 1969), and phylogenetic trees were obtained by neighbor-joining (Saitou & Nei, 1987). All positions containing alignment gaps and missing data were eliminated in pairwise sequence comparisons (pairwise deletion option). Stability of clades was assessed with 1000 bootstrap replications (Felsenstein, 1985).

The selected isolate was also examined for several physiological properties using standard yeast identification methods, according to Kurtzman & Fell (1998). Assimilation of saccharose, fructose, rhamnose, trehalose, maltose, glycerol, melezitose, inulin, melibiose, inositol, lactose, raffinose, galactose, nitrate, and creatine was assessed. Carbohydrate and glycerol assimilation was carried out in yeast nitrogen base (Difco), while nitrate and creatine assimilation was assessed in yeast carbon base (Difco). Fermentation of glucose, urea hydrolysis, and the diazonium blue B reaction were also determined as described by Kurtzman & Fell (1998). Growth at different temperatures was assessed in 50-mL Erlenmeyer flasks containing 10 mL of yeast nitrogen broth (YNB; Difco) amended with 0.5% of glucose in orbital shaker at 150 r.p.m.

Biofilm formation

The ability to form a biofilm was assessed using a modification of the adherence assay described by Ruzicka et al. (2007) for Candida parapsilopsis. Briefly, cells from 2-day-old PDA cultures were suspended in sterile water to reach a turbidity corresponding to a value of 4 of the McFarland scale. Wells of a 96-well microtiter plates (Montegrotto Terme, Padova, Italy) containing 180 μL of sterile apple juice were inoculated with 20 μL of a yeast suspension and incubated for 48 h at 25 °C. Negative controls containing only sterile apple juice were included. After incubation, the wells were emptied, rinsed with water, and air-dried at room temperature. The adherent biofilm layer was stained with an aqueous solution of violet crystal 1% (w/v) for 20 min, rinsed with water, and air-dried. The bound dye was eluted from each well with 200 μL of ethanol. The absorbance of each well was measured at 620 nm. Each treatment was repeated three times in the same plate. Biofilm formation was considered positive in those wells where absorbance was higher than the mean of the negative control plus three standard deviations (Ruzicka et al., 2007).

Siderophore production

Siderophore production was assessed with a chrome azurol sulfonate (CAS) agar plate assay (Schwyn & Neilands, 1987) using Grimm Allen medium as a base (Baakza et al., 2004). Each isolate was streaked in a section of the plate and incubated at 25 °C for 5 days. After incubation, the presence of a siderophore was indicated by a color change from blue to orange surrounding the streaks of growing cultures.

Sensitivity of At17 to fungicides and diphenylamine

Sensitivity of At17 to several fungicides was assessed. The minimal inhibitory concentration (MIC) of iprodione, thiabendazole, and imazalil was determined on PDA amended with different fungicide concentrations. Cultures of At17 were prepared as described. The suspension was adjusted by dilution to 104 CFU mL−1, and 10 μL of suspension was inoculated onto 200 μL of PDB amended with different concentrations of fungicide dispensed into the wells of sterile, disposable, 96-well microtiter plates (Montegrotto Terme). After 72 h of incubation at 25 °C in darkness, microbial growth was determined visually. The MIC was defined as the lowest concentration that inhibited microbial growth. Fungicide concentrations assayed were 0, 1, 2, 4, 8, 16, 32, 64, and 128 µg mL−1. Formulated fungicides were used for these experiments. Rovral 50 WP (Rhône Poulenc, Lyon, France), Tecto 500SC (500 g L−1; Syngenta International AG, Basel, Switzerland), and Fungaflor 500EC (Janssen Pharmaceutica N.V., Bélgica) were used for iprodione, thiabendazole, and imazalil, respectively. The amount of formulated fungicide added to the medium was calculated to reach the concentrations of active ingredients specified previously. The experiment was repeated three times.

The ability of At17 to stay viable in water containing commercial concentrations of diphenylamine was also assessed. Cells of At17 (10 μL of a suspension containing 1 × 107 CFU mL−1 in sterile distilled water) were inoculated in 10-mL tubes containing 5 mL of sterile water amended with 0 or 1000 μg mL−1 of diphenylamine. The concentration of viable cells after 20, 30, 60, and 120 min of incubation was assessed by plate count on PDA.

Wound colonization by the selected antagonist in the absence and presence of diphenylamine

Growth curves for At17 in wounds of apple stored at 1 °C were constructed in the absence and presence of diphenylamine applied at a commercial concentration. Apples were wounded as described. Wounds were then inoculated with 10 μL of a yeast cell suspension (1 × 107 CFU mL−1) in either sterile water or diphenylamine (1000 µg mL−1) solution. After inoculation, fruits were placed in boxes at 1 °C. At each collection time (0, 2, 24, 96, 168, 218, 360, 480 h, and 3 months), the surrounding tissue from the wounds of three different apples was removed and placed in a 1.5-mL Eppendorf tube containing 1 mL of aqueous sterile solution of Tween 20 (0.1%). Tubes were homogenized in a vortexer (Tecnolab, SP, Brazil) for 2 min. Quantification of viable cells in each sample was performed by plate count on PDA. For each antagonist, growth curves were constructed by plotting the logarithm of the number of viable cells in each wound versus time of storage. Exponential growth phase was determined in each case.

Biocontrol assay in the presence of diphenylamine

An experiment was conducted to assess the biocontrol activity of At17 in the presence of commercial levels of diphenylamine. The test was conducted on ‘Red Delicious’ apples as previously described. Briefly, wounds were inoculated with a 10 μL of a 1 × 107 CFU mL−1 suspension of At17 in diphenylamine (1000 μg mL−1). Inoculation of apples with At17 suspended in water served as a control. Following inoculation, apples were stored for 3 months at 0–1 °C at which time the apples were removed and disease incidence recorded.

Mode of action studies

In vitro antagonism

The ability of At17 to inhibit growth of P. expansum and B. cinerea in vitro was tested using dual cultures on apple juice agar as described by Spadaro et al. (2002). Briefly, a line of the antagonist was plated on the medium using cells collected from 2-day-old cultures. Ninety mm Petri dishes were used, and the strip of antagonist was c. 25 mm from the border. A 5-mm mycelial disk of the pathogen taken from the edge of colonies growing on Petri dishes containing PDA was placed 30 mm from the border and 30 mm from the antagonist. Plates were incubated at 25 °C until the mycelium of the pathogen reached the Petri dish border opposite to the antagonist strip. The radial growth of the pathogen toward the antagonist was then measured and was compared with the opposite colony radius (30 mm). Experiments were repeated three times.

Competition for iron

Dual cultures of At17 with native strains of P. expansum and B. cinerea in the presence and absence of iron (0 and 500 μM of ferric chloride) were plated. The medium used in this assay was the same as described by Calvente et al. (1999).

Production of antifungal volatile compounds

Production of volatile antibiotics was tested as described by Huang et al. (2011) with modifications. Plates containing PDA were inoculated centrally with an agar plug containing P. expansum or B. cinerea. At the same time, another plate containing 10 mL of PDA was superficially inoculated with 100 μL of a suspension of 1 × 107 CFU mL−1 of At17, which was evenly spread on the surface of PDA with a sterilized stainless steel spatula. The covers of inoculated plates were removed, and the PDA dishes inoculated with the yeast were covered above the dishes inoculated with P. expansum or B. cinerea. The sets of double dishes were sealed using double layers of Parafilm to make closed chambers. Control plates were inoculated with pathogens but covered with PDA plates not inoculated with yeast. The closed plates were incubated at 25 °C for 3 days. The pathogen colony diameter was measured after incubation. Inhibition was expressed as percentage of colony diameter reduction compared with the control.

Effect of At17 on pathogen spore germination

The effect of At17 antagonist on spore germination of P. expansum and B. cinerea was assessed in sterile apple juice as described by Spadaro et al. (2002). Aliquots (10 μL) of spore suspension (1 × 106 spores mL−1) of each pathogen in sterile distilled water were transferred to the wells of sterile, disposable, 96-well microtiter plates (Montegrotto Terme), containing 100 μL of sterile apple juice and 10 μL of a suspension containing 1 × 107 cells mL−1 of At17. Wells containing pathogen spores without the addition of At17 cells served as controls. Each experiment consisted of three replicates (individual wells), and the experiment was repeated three times. After 12-h incubation at 25 °C, 100 spores per replicate were observed microscopically, and their germination was evaluated. Five to ten fields within each well were selected randomly to obtain the germination data. Data for percentage of germinated spores (spore germination) were transformed into arcsine square root values to normalize distribution prior to conducting an analysis of variance. Mean separations were performed using Duncan's multiple range test as described in Zheng et al. (2004).

Production of antifungal enzymes

Production of chitinase and β 1, 3-glucanase by At17 was assayed in yeast nitrogen base (YNB) in the presence of fungal cell walls as described by Vero et al. (2009). At different times (3, 7, 20, 25, 28, and 30 days), aliquots from the three different cultures were filtered through a 0.45-μm filter, and enzymatic activities were assayed in the filtrates. Chitinase activity was assayed as described by Mahadevan & Crawford (1997) by measuring the release of p-nitrophenol from p-nitrophenyl-N, acetylglucosamidine [pNP(GlcNAc)]. One unit of enzyme was defined as the amount releasing 1 μmol of paranitrophenol per mg protein.

β 1,3-Glucanase was assayed as in Masih & Paul (2002), by determining the amount of reducing sugars released from laminarin, according to the Nelson–Somogyi method (Nelson, 1944; Somogyi, 1952) using glucose as a standard. One unit of β 1,3-glucanase was defined as the amount of enzyme releasing 1 μmol of reducing sugars per mg protein.

Statistical analysis

Incidence of rot was analyzed using the generalized linear model, assuming a binomial distribution, with a logit transformation (maximum likelihood). Models were adjusted by overdispersion, and variance due to fruits and trays were studied and estimated. The Glimmix procedure of the sas statistical package (9.1.3 release) was used (SAS institute, Cary, NC, 2005).

Data on biofilm formation and effect of At17 on pathogen spore germination were analyzed using anova, and the means were compared by the LSD test at a significance level of 0.05 using the infostat software package (Grupo InfoStat, FCA, Universidad Nacional de Córdoba, Argentina, 2009).

Results

Isolation and identification of potential biocontrol agents

Only five yeast isolates were obtained from the Antarctic soil samples using the described culture conditions. Table 1 shows the species identified by molecular analysis. Both ITS and D1/D2 consensus sequences exhibited ≥ 99% homology to corresponding sequences deposited in GenBank for the respective species. The same species designation was obtained for each isolate when the blast analysis was conducted with either the obtained ITS sequence or the obtained D1/D2 sequence.

Table 1. Antarctic yeast isolates used in this study
IsolateSpecies
At 1Cryptococcus terricola, Pedersen
At 4Cryptococcus gastricus, Reiersol & di Menna
At 7Rhodotorula mucilaginosa, (Jörgensen) F.C. Harrison
At 16Rhodotorula laryngis, Reiersöl
At 17Leucosporidium scottii, Fell, Statzell, Hunter & Phaff

Selection of biocontrol agents

All the isolates obtained from the Antarctica soil samples, as well as PL1 (a previously selected yeast antagonist), significantly reduced the percentage of decayed wounds in ‘Red Delicious’ apples resulting from P. expansum and B. cinerea when compared with the control (Fig. 1a and b). However, isolates PL1 (C. infirmominiatum) and At17 (L. scottii) showed the highest level of protection against B. cinerea. When tested against P. expansum on ‘Pink Lady’ apples, protection levels achieved with PL1 and At17 were again significantly higher than those obtained with the other isolates (Fig. 1c). As a result, At17 was selected for further study.

Figure 1.

Incidence of diseases on apples inoculated with different yeast isolates and pathogens. At1, At4, At7, At16, and At17 represent Antarctic isolates; PL1 is a previously selected antagonist. Controls were not treated with yeast. Fruit was held for 3 months at 0–1 °C. Treatments with the same letter are not significantly different (= 0.05). (a) Incidence of blue mold on ‘Red Delicious’ apples inoculated with different yeast isolates and Penicillium expansum. (b) Incidence of gray mold on ‘Red Delicious’ apples inoculated with different yeast isolates and Botrytis cinerea. (c) Incidence of blue mold on ‘Pink Lady’ apples inoculated with different yeast isolates and P. expansum. (d) Incidence of blue mold on ‘Red Delicious’ apples inoculated with PL1 and At17 as a mixture and separately.

PL1 and At17 were also tried as a mixture to protect wounds in ‘Red Delicious’ apples against P. expansum (Fig. 1d). Disease incidence with the 50/50 (v/v) mixture was significantly lower than the control (24% vs. 99%, respectively). However, use of the mixture did not increase the level of control over the level obtained with the use of each isolate separately.

Phylogeny of the selected antagonist

Phylogenetic analysis of At17 utilizing both the D1/D2 and the ITS1-ITS2 domains of rDNA sequences within a group of sequences representing different species belonging to the order Leucosporidiales placed At17 in the L. scottii clade closely related to Leucosporidiella creatinivora. Only phylogenetic tree derived from analysis of 26S rDNA domain D1/D2 is shown in Fig. 2 as the tree obtained by analyzing ITS1-ITS2 domain showed essentially the same relationships among isolates and species. ITS1-ITS2 sequence from At17 differed by one nucleotide with the corresponding sequence from L. scottii CBS 614 type strain, and by three and four nucleotides with sequences corresponding to L. creatinivora type strain CBS 8620 and L. scottii CBS 5930 type strain, respectively. However, D1/D2 sequence from At17 showed a 100% homology with the same sequences from L. scottii CBS 614 and CBS 5930 type strains and differed by one substitution with the sequence corresponding to L. creatinivora type strain CBS 8620. The nucleotide sequences corresponding to the D1/D2 and the ITS1-ITS2 domains of rDNA of At17 have been deposited in GenBank under the following accession numbers: JX296532 and JX296533.

Figure 2.

Phylogenetic tree showing placement of At17 among species of the order Leucosporidiales and reference species Rhodotorula glutinis and Rhodotorula graminis (the outgroup species in the analysis) as represented by the bootstrap consensus tree inferred from 1000 replicates derived from analysis of 26S rDNA domain D1/D2. The percentages of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Jukes–Cantor method and are in the units of the number of base substitutions per site. All positions containing gaps and missing data were eliminated. There were a total of 518 positions in the final data set. All but At17 are type strains.

Phenotypic characterization of the selected antagonist

A compilation of biochemical characteristics is given in Table 2. Biochemical characteristics are coincident with those described for L. scottii (Fell et al., 1969) and differed from other species of the genus. Results obtained for rhamnose (+) and creatine (−) assimilation allowed the differentiation with L. creatinivora, the closest specie in both phylogenetic trees. Based on both the molecular and phenotypic characterization, At17 was identified as L. scottii.

Table 2. Results of physiological and biochemical tests of At17
Test AssimilationResultTest AssimilationResult
Glucose+Inuline
Saccharose+Melibiose
Fructose+Inositol
Rhamnose+Lactose
Trehalose+Raffinose
Maltose+Galactose
Glycerol+Nitrate+
MelizitoseCreatine+
Fermentation Growth
Glucose25 °C+
37 °C
Additional tests Without vitamins+
Urea hydrolysis+  
Diazonium blue B reaction+  

Biofilm formation

The threshold value for biofilm formation was set at A620 = 0.322, based on absorbance values obtained for negative controls (Ruzicka et al., 2007). Table 3 lists the mean absorbance, confidence intervals (α = 0.05), and standard deviation (SD) for the two yeast species represented by the two isolates and controls (empty wells). The data indicate that L. scottii, strain At17, formed a biofilm in the assayed conditions, while C. infirmominiatum, strain PL1, did not.

Table 3. Assessment of biofilm formation in yeast isolates (PL1 and At17) as measured by the mean absorbance value of solutions eluted from each sample after they had been stained with crystal violet
TreatmentA620 (mean ± SD)
  1. Increasing levels of absorbance reflect the ability of the isolate to form a biofilm. Each value represents a mean ± standard deviation (SD). Treatments with the same letter are not significantly different (= 0.05).

Control0.256 ± 0.022a
PL10.265 ± 0.044a
At170.661 ± 0.105b

Siderophore production

An orange halo of about 2 cm was formed around the streak of At17 when grown on CAS medium indicating siderophore production. Smaller orange halos (about 1–2 mm) were present around streaks of PL1, Rhodotorula mucilaginosa, strain At7, and Rhodotorula laryngis, strain, At16, suggesting that siderophore production was minimal (Fig. 3).

Figure 3.

Siderophore production on CAS plates. Orange halos around streaks indicate siderophore production.

Sensitivity of At17 to fungicides and diphenyl amine

Growth of At17 was observed in the presence of the highest concentration of all the tested fungicides. Therefore, it can be assumed that the selected antagonist is resistant to 128 mg mL−1 of iprodione, thiabendazole, and imazalil in vitro. Moreover, viability of At17 cells was not affected after 2 h of incubation in an aqueous solution of 1000 μg mL−1 diphenylamine (Table 4), which is the current recommended concentration for commercial use.

Table 4. Survival of At17 following exposure to 1000 p.p.m. diphenylamine in water for up to 2 h
Time (h)Log viable cell mL−1 (mean ± SD)
  1. Data are expressed as the log of the mean concentration of viable cells (CFU mL−1). Each value represents a mean ± standard deviation (SD). No statistically significant differences were observed.

  2. a

    Each value is the mean of three experiments.

04.7 ± 0.2a
0.54.8 ± 0.3
14.7 ± 0.2
24.5 ± 0.3

Wound colonization by selected antagonist

Growth curves of At17 for a period of 20 days following inoculation into apple wounds, in the absence and presence of diphenylamine, are shown in Fig. 4. Differences in the growth profiles were not detected. In both cases, the number of viable cells decreased after the yeasts were administered into apple wounds. Exponential growth was initiated by 24 h postinoculation, and the yeast exhibited a growth rate of 0.012 h−1 in both cases. A lag phase was not detected in either case, and stationary phase was reached by the 5th day. In both cases, viable cells remained constant for a period of 3 months (data not shown).

Figure 4.

Growth curves of At17 in apple wounds at 1 °C, in the presence and in the absence of 1000 μg mL−1 of diphenylamine (DPA). Bars represent confidence intervals (α = 0.05).

Biocontrol assay in presence of diphenylamine

No significant differences in wound protection against P. expansum were detected when the antagonist was applied in the presence of diphenylamine. Incidence of blue mold in At17-treated wounds in the presence or absence of diphenylamine was 11% in each case, while it reached 99% in control apples (data not shown).

Mechanisms of biocontrol

In vitro antagonism

Pencillium expansum radial growth was inhibited by 50% when dual cultures were grown on apple juice agar. The level of inhibition of radial growth of P. expansum was the same in the presence or absence of iron, and inhibition was maintained in the presence of 500 μM ferric chloride. No inhibition of B. cinerea growth in vitro was observed under any of the experimental conditions.

Production of antifungal volatile compounds

A reduction of 20% in the rate of colony growth of both pathogens was observed in the presence of volatiles produced by At17 (data not shown). The effect of the volatiles was only inhibitive as both pathogens continued to grow until they reached the edge of the plate.

Effect of At17 on pathogen spore germination

Spore germination of both pathogens in apple juice was significantly reduced in the presence of living cells of At17. An inhibition of 71% and 63% was determined for P. expansum and B. cinerea, respectively (Table 5).

Table 5. Influence of At17 on the germination of Penicillium expansum and Botrytis cinerea spores in apple juice. Control represents germination of spores in apple juice in the absence of At17
TreatmentsSpore germination (%)
Penicillium expansum Botrytis cinerea
  1. a

    Each value is the mean of three experiments. Values in the same column with the same letter are not significantly different (= 0.05).

At1729aa37a
Control98b95b

Production of antifungal enzymes

At17 exhibited minimal growth using P. expansum cell walls as the only carbon source. Furthermore, the enzymatic assays indicated that the AT17 did not produce significant amounts of either β-glucanase or chitinase under the assayed conditions (data not shown).

Discussion

The strategy used in this study proved to be an efficient and reliable approach to isolate yeasts that could serve as potential biocontrol agents to control postharvest diseases of apples, and perhaps other fruit, stored at low temperatures. It consisted of a selective enrichment of microflora in apple juice followed by isolation and purification on apple juice agar, all conducted at 8–10 °C. This procedure favored the selection of several psychrotrophic microorganisms able to colonize apple wounds. Importantly, this selection strategy minimized the number of isolates to be evaluated in larger-scale fruit studies conducted at low temperature. Thus, it allowed the selection of potential antagonists with a minimal expenditure of time and expense.

Only five yeast species/isolates were obtained. All of them belonged to the phylum Basidiomycota and were identified as belonging to three genera: Cryptococcus, Rhodotorula, and Leucosporidium. The fact that all our species were in the Basidiomycota may be a reflection of the findings of Connell et al. (2008) who studied the diversity of soil yeasts from South Victoria lands, Antarctica, and found that 89% of isolates were basidiomycetous species. They reported that Cryptococcus represented the genus most often isolated followed by Leucosporidium. In their report, they identified the same Rhodotorula species (R. mucilaginosa and R. laryngis) that were isolated in this study. Rhodotorula and Cryptococcus were also among the most frequently isolated genera of cold-adapted yeasts identified from glacial melt water rivers in Patagonia, Argentina (De García et al., 2007).

All the tested yeast isolates significantly reduced the percentage of decayed wounds in ‘Red Delicious’ apples caused by P. expansum and B. cinerea. Only L. scottii (At17), however, provided significant protection of ‘Pink Lady’ apples. Differences in biocontrol efficacy in the different apple cultivars may be due to differences in the pH of the pulp, which was 4.0 in ‘Red Delicious’ apples and 3.5 in ‘Pink’ Lady apples. Leucosporidium scottii had been previously described as a biocontrol agent of postharvest diseases of apples and citrus (Vero et al., 2011). In that work, it had been isolated from cold-stored lemons.

Leucosporidium scottii is an aerobic, basidiomycetous yeast that has been isolated from low-temperature environments, generally in polar regions or temperate regions during cold weather seasons (Kurtzman & Fell, 1998). Its presence in Antarctic soils and water has been previously reported by Fell et al. (1969). Leucosporidium scottii is the type species of the genus, which includes six species (Sampaio, 2011). Based on D1/D2 and ITS sequence data, the genus is currently polyphyletic, with two species (Leucosporidium antarcticum and Leucosporidium fasciculatum) falling outside the recently described order Leucosporidiales (Turchetti et al., 2011). In fact, phylogenetically, L. scottii is more related to L. creatinivora than to other species of the genus. In the present study, phenotypic characterization of the selected yeast was used specifically to differentiate between the two species.

Leucosporidium scottii is considered a psychrotrophic yeast, differing in its maximal growth temperature from other species of the genus. It can grow at 30 °C, while other species of the genus only grow at lower temperatures (25 °C; Kurtzman & Fell, 1998). The At17 isolate of L. scottii has an optimal growth temperature of 28 °C, which makes it economically feasible to be produced in large-scale fermentors without additional costs needed to produce cooler fermentor conditions. Otherwise, it is unable to grow at human body temperature (37 °C), which is a very advantageous characteristic when selecting a biocontrol agent for application to fruit in a postharvest environment that will later be ingested by humans (Wilson and Wisniewski, 1994; Wisniewski et al., 2007).

Another important characteristic of At17 was its ability to form biofilms when grown in apple juice. Biofilm formation is a process in which microorganisms form multicellular structures embedded in a complex matrix that allows them to improve adhesion to surfaces and to enhance resistance to stresses. It is considered an important attribute of postharvest antagonists, which assists in the ability of the yeast antagonist to successfully colonize and protect both wounds and intact fruit surfaces (Droby et al., 2009).

The biocontrol activity of At17 was compared with a previously selected yeast antagonist, PL1, identified as C. infirmominiatum (Vero et al., 2011). Both PL1 and At17 are aerobic microorganisms and are able to assimilate the three main monosaccharides (glucose, sucrose, and fructose) present in apple juice (Vero et al., 2002). Both isolates demonstrated that they could grow in apple wounds at low temperatures and remain viable for at least 3 months.

At17 was resistant to the three fungicides most frequently used in postharvest applications in Uruguay. Resistance to commonly used postharvest has been previously reported for Meyerozyma (Pichia) guilliermondii (Droby et al., 1993) and for Candida saitoana (El Ghaouth et al., 2000), so these results are not unexpected. This implies that isolate At17 could be applied to fruit in combination with these fungicides as part of an integrated management practice without affecting the viability of the antagonist. At17 also maintained viability for at least 2 h when exposed to diphenylamine, a chemical used to prevent apple scald. Additionally, the growth of At17 in apple wounds was not inhibited or otherwise affected despite the presence of diphenylamine. This result indicated that At17 could be applied in aqueous suspension together with diphenylamine, thus avoiding the addition of an additional step to the fruit processing line to accommodate the application of the yeast by itself.

At17 produced soluble antifungal substances that were inhibitory to P. expansum but not B. cinerea when the organisms were grown in dual cultures in three different media. At17 also produced siderophores when grown in CAS medium. However, such compounds did not seem to play a role in pathogen inhibition as the amount of radial growth of P. expansum was the same regardless of the presence or absence of iron, and no inhibition of B. cinerea was observed in vitro under any of the experimental conditions.

The production of volatile compounds by At17 having inhibitory antifungal activity against both pathogens was observed. Inhibition of fungal pathogens by volatile compounds produced by yeasts has been previously reported (Druvefors et al., 2005; Gholamnejad et al., 2009; Huang et al., 2011). Druvefors et al. (2005) identified ethyl acetate and ethanol as the volatile compounds involved in the inhibition of Penicillium roquefortii by Wickerhamomyces anomalus (Pichia anomala), while Huang et al. (2011) found that 1,3,5,7-cyclooctatetraene and 3-methyl-1-butanol, were the most abundant volatile compounds with antifungal activity against B. cinerea, produced by Candida intermedia.

At17 significantly inhibited P. expansum and B. cinerea spore germination in apple juice. This observation has been reported for many biocontrol yeasts, including those that inhibited mycelial growth (Tongsri & Sangchote, 2009; Hernández-Montiel et al., 2010) and those that did not (Vero et al., 2009). This result was not surprising as the production of soluble and volatile antifungal compounds by At17 was demonstrated. Therefore, it could be hypothesized that such compounds could be involved in spore germination inhibition. The soluble antifungal compounds, however, showed specific activity against P. expansum and not B. cinerea mycelial growth. Production of fungal cell wall–degrading enzymes by At17 could not be detected even when it was grown in the presence of P. expansum cell walls as the only carbon source. Further studies are needed to evaluate the full potential of this newly identified yeast antagonist as a postharvest biocontrol agent during cold storage. While other environments, such as chilled and frozen food product, have served as a source for psychrotrophic microorganisms, this study suggests that yeast isolated from polar regions may also be an excellent source for obtaining novel biocontrol agents that need to function in cold environments such as those found during the cold storage of fruits for extended periods of time.

Acknowledgements

We are very grateful to Instituto Antártico Uruguayo (IAU) who supported this work. We also want to thank Dr Silvia Batista for providing Antarctic soil samples.

Ancillary