Biological Control
Improving Biological Control by Combining Biocontrol Agents Each with Several Mechanisms of Disease Suppression Ruth Guetsky, D. Shtienberg, Y. Elad, E. Fischer, and A. Dinoor First, second, and third authors: Department of Plant Pathology, ARO, The Volcani Center, Bet Dagan 50250; fourth author: Institute of Soil and Water, ARO, The Volcani Center, Bet Dagan 50250; and first and fifth authors: Department of Plant Pathology and Microbiology, The Hebrew University of Jerusalem, Faculty of Agricultural, Food and Environmental Quality Sciences, Rehovot 76100 Israel. Accepted for publication 28 April 2002.
ABSTRACT Guetsky, R., Shtienberg, D., Elad, Y., Fischer, E., and Dinoor, A. 2002. Improving biological control by combining biocontrol agents each with several mechanisms of disease suppression. Phytopathology 92:976-985. Two biocontrol agents, a yeast (Pichia guilermondii) and a bacterium (Bacillus mycoides), were tested separately and together for suppression of Botrytis cinerea on strawberry leaves and plants. Scanning electron microscopy revealed significant inhibition of Botrytis cinerea conidial germination in the presence of Pichia guilermondii, whereas Bacillus mycoides caused breakage and destruction of conidia. When both biocontrol agents were applied in a mixture, conidial destruction was more severe. The modes of action of each of the biocontrol agents were elucidated and the relative quantitative contribution of each mechanism to suppression of Botrytis cinerea was estimated using multiple regression with dummy variables. The improvement in control efficacy achieved by introducing one or more mechanisms at a time was calculated. Pichia guilermondii competed with Botrytis cinerea for glucose, sucrose, ade-
Biological control is a nonchemical measure that has been reported in several cases to be as effective as chemical control (10, 14). However, the efficacy of biological control is occasionally inadequate and variability in control efficacy may be high. Understanding the mechanisms involved in biological control may enable enhancing control efficacy and reducing the inconsistency and variability. The mechanisms involved in biological control are several and include, among others, induced resistance, competition for nutrients, and secretion of inhibitory compounds. Induced resistance, or as it is often referred to, induced systemic resistance, is a complex biocontrol process consisting of various mechanisms. It may be activated by chemical compounds (elicitor molecules) or by some nonpathogenic microorganisms (9,19). Several types of elicitor molecules are known, such as salicylic acid, complex carbohydrates, fatty acids, amino acids, glycoproteins, yeast-derived elicitors, and microbial metabolites (22). Salicylic acid, for example, was identified as a secondary signal that is induced by a primary translocated signal generated at the infection site (28). Activation of the defense mechanisms may be rapid. Eight hours after inoculation of a cucumber leaf with Pseudomonas syringae pv. syringae, salicylic acid was detected in phloem exudates of that leaf; 12 h after inoculation, it was detected in the phloem of a leaf located above the inoculation site (35). Nevertheless, the control efficacy achieved by this mechanism is highly Corresponding author: D. Shtienberg; E-mail address:
[email protected] Publication no. P-2002-0718-01R This article is in the public domain and not copyrightable. It may be freely reprinted with customary crediting of the source. The American Phytopathological Society, 2002.
976
PHYTOPATHOLOGY
nine, histidine, and folic acid. Viability of the yeast cells played a crucial role in suppression of Botrytis cinerea and they secreted an inhibitory compound that had an acropetal effect and was not volatile. Bacillus mycoides did not compete for any of the sugars, amino acids, or vitamins examined at a level that would affect Botrytis cinerea development. Viable cells and the compounds secreted by them contributed similarly to Botrytis cinerea suppression. The bacteria secreted volatile and nonvolatile inhibitory compounds and activated the defense systems of the host. The nonvolatile compounds had both acropetal and basipetal effects. Mixture of Pichia guilermondii and Bacillus mycoides resulted in additive activity compared with their separate application. The combined activity was due to the summation of biocontrol mechanisms of both agents. This work provides a theoretical explanation for our previous findings of reduced disease control variability with a mixture of Pichia guilermondii and Bacillus mycoides. Additional keyword: gray mold.
variable. For example, Trichoderma harzianum T39 induces plant defense against Botrytis cinerea in tomato, lettuce, pepper, bean, and tobacco. Application of T. harzianum T39 at a site spatially separate from inoculation with Botrytis cinerea results in a 25 to 100% reduction in disease severity (9). Nutrients are essential for the development of populations of epiphytic microorganisms, necrotroph pathogens, and nonpathogens alike. Competition for nutrients on plant surfaces is an important mechanism of biological control against pathogens that depend on external nutrition (5,12,25,29). For example, isolates of Rhodotorula glutinis and Cryptococcus albidus compete for nutrients with germinating conidia of Botrytis cinerea (5,11,13). Similarly, several bacterial isolates compete for glucose and asparagine with germinating oospores of Pythium aphanidermatum in the rhizosphere of various crops (12), and Pseudomonas fluorescens suppresses Bipolaris maydis by competing with the pathogen for glucose (24). Biological control that is based on competition for limited nutrients can be nullified by increasing the concentrations of the relevant nutrients (5,11,13). The degree of disease suppression achieved by competition is governed by numerous biotic (e.g., host species and nutritional status) and abiotic (e.g., temperature and relative humidity) factors and its efficacy cannot be predicted accurately. Volatile and nonvolatile compounds produced by microorganisms may be involved in the suppression of plant pathogens as well (16,20,24,27,34). For example, Pseudomonas fluorescens and Bacillus pumilus inhibit Botrytis cinerea on strawberries by secreting volatile compounds with fungistatic effects (33). Involvement of nonvolatile antibiosis has been described in many systems in which bacterial isolates were applied as biocontrol agents (16,34). Bacillus subtilus controlled Monilia fructicola on stone fruits by secreting antifungal peptides, identified as iturin
antibiotics. This antibiotic substance was equal to benomyl in efficacy (27). Similarly, Botrytis cinerea was suppressed by Pseudomonas antimicrobica in vitro and on strawberry leaves. Antifungal compounds were detected in cell-free filtrates of the bacterium (34). In most studies, the involvement of only one mechanism of biological control is demonstrated. Involvement of more than one mechanism has been reported in only a few systems. Those reports include antibiosis together with enzyme degradation of Botrytis cinerea cell wall and competition for nutrients, followed by interference with pathogenicity enzymes of the pathogen or induced resistance (11). A commercial product consisting of a mixture of three Bacillus subtilis strains was used to control fungal soil pathogens after disinfection by seed-treating fungicides. These strains exhibit several biocontrol mechanisms, including production of antifungal compounds (including antibiotics and hydrogen cyanide), competition for ferric iron, competition for infection sites, and production of lytic enzymes (27). Biocontrol agents are affected by biotic and abiotic conditions. Because different mechanisms of control may be dissimilarly influenced by those conditions, it is possible that if multiple mechanisms are involved, under a certain set of conditions one mechanism may compensate for the other. Therefore, we hypothesize that control efficacy achieved by biocontrol agents exhibiting several distinct mechanisms of control will result in additive or synergistic, but not antagonistic, effects. Biological control with multimechanisms may be achieved by using one biocontrol agent exhibiting several mechanisms or by applying more than one biocontrol agent in a mixture, provided that each of them has one (or several) distinct mechanisms. We tested these hypotheses using the gray mold–strawberry pathosystem and two biocontrol agents, a bacterium and a yeast, as a model. The gray mold disease caused by Botrytis cinerea Pers.; Fr. inflicts serious losses in many crops (31). In strawberry, the fungus attacks flowers, setting fruits, mature fruits, and leaves (6,18,31, 32). The main sources of inoculum for the disease in strawberry are mummified fruits, dead leaves, straw mulch (where used), and neighboring crops (18,32). Infected flower parts shed after bloom and adhere to the fruit surface. After that, the pathogen may cause quiescent infection until the fruits ripen (6,32). Diseased fruits are mostly covered by gray tuft composed of mycelia, conidiophores, and conidia of the fungus, and eventually, the fruits rot. Bacillus mycoides is a gram-positive rod-shaped bacterium containing an endospore and relatively thick cell wall (7). Bacillus strains have been the most frequently exploited bacteria for commercial development of biocontrol agents because of their resistant endospores, which may remain viable for long periods and are tolerant to extreme temperatures and pHs (2). Pichia guilermondii Wick. is a white yeast with vegetative multilateral budding. When asci are formed, they contain one to four hat-shaped ascospores (3). Both biocontrol agents were selected from several microbial candidates based on their performance as biocontrol agents of Botrytis cinerea in strawberry. The two biocontrol agents differ in their ecological requirements. Control efficacy achieved by a mixture of Pichia guilermondii and Bacillus mycoides was, in some cases, higher than that achieved by their separate application. Moreover, when the mixture was used, the inconsistency and the variability of Botrytis cinerea suppression were reduced under a range of environmental conditions (15). In the present study, the specific mechanisms involved in the biological control of Botrytis cinerea by Pichia guilermondii and Bacillus mycoides were investigated. In experiments conducted under controlled conditions, we estimated the relative quantitative contribution of each mechanism to the suppression of Botrytis cinerea, and the improvement in control efficacy achieved by introducing one (or more) mechanisms at a time. In addition, we examined the disease control efficacy achieved by applying both biocontrol agents in a mixture to the strawberry phyllosphere.
MATERIALS AND METHODS Organisms. Botrytis cinerea was cultured on potato dextrose agar (PDA, Difco Laboratories, Detroit) in petri plates and incubated at 20°C. Conidia were harvested from 10- to 14-day-old cultures by agitating small pieces of agar bearing mycelia and conidia in a glass tube containing 2 ml of tap water and 0.01% (wt/vol) Tween 80. The suspension was filtered through cheesecloth, and the spore concentration was calibrated with a hemacytometer and adjusted to 1 × 106 spores per ml. Two biocontrol agents, a yeast and a bacterium, previously isolated from tomato leaves (15), were used in all the trials. Pichia guilermondii (isolate Y2) was grown on PDA or on Luria-Bertani (LB; solid or liquid) media for 24 to 48 h at 25°C before use. Bacillus mycoides (isolate B16) was grown on solid or liquid nutrient media (Difco) for 24 h at 30°C. The same medium was used in any particular experiment, for all the treatments. The yeast and the bacterial cells were washed from the agar media in 10 ml of saline solution (8 mg of NaCl per ml) supplemented with 0.01% Tween 80. Cell concentrations were determined and adjusted to 1 × 107 CFU/ml. Strawberry plants (Fragaria ananassa cv. OsoGranade) were used in all experiments. Cultivar Oso-Granade is highly susceptible to Botrytis cinerea. Seedlings were planted at the beginning of September each year in plastic pots (containing growth mixture with 70% peat and 30% vermiculite) and maintained in a greenhouse under temperatures of 20 to 30°C. Plants were irrigated and fertilized (70, 30, 70, and 100 ppm of N, P, K, and phosphoric acid, respectively) weekly as needed to maintain full moisture capacity in the growth medium. Inoculation and disease assessment. Unless otherwise stated, treatment effects were determined on detached strawberry leaflets. Drops (20 µl) containing Botrytis cinerea conidia were mixed, or not, with the biocontrol agents (1:1, vol/vol) and placed on detached strawberry leaflets (5 drops per leaflet). Glucose (0.05 or 0.1%, wt/vol, final concentration) and KH2PO4 (0.05%, wt/vol, final concentration) were added to the spore suspension. The leaflets were then put on a plastic net in boxes (38 × 32 × 17 cm, length × width × height) filled with 500 ml of distilled water. The boxes were covered with polyethylene bags to maintain high relative humidity and placed in a growth chamber at 20°C. Germination of Botrytis cinerea conidia was determined after 26 h of incubation, and disease severity was recorded after 7 days. Germination of Botrytis cinerea conidia was determined as follows: pieces of strawberry leaflets bearing a drop of the interacting microorganisms were placed on glass slides, stained with aniline blue, incubated at room temperature (20 to 25°C) for 15 min, and observed under a light microscope. Germination was determined for 50 conidia on each leaflet. Disease severity was determined on the detached leaflets, as follows: the diameter of the 20-µl drop on a strawberry leaflet was 4 to 5 mm. Under the controlled conditions used in this study, the diameter of Botrytis cinerea lesions formed within 7 days was approximately 9 to 12 mm. A 12-mm-diameter lesion was used as a base size and assigned a value of 100%. A pictorial scale of lesion sizes was then prepared, including the following relative sizes: 0, 0.5, 1, 2, 5, 10, 20, 40, 75, and 100%. Lesion size (a measure of disease severity) was determined for each lesion separately using the pictorial scale. Control efficacy (%CE) was determined for each treatment by using the numbers of germinating conidia and the estimates of disease severity, in treated and untreated leaflets (Dt and Du, respectively), in the following formula: CE = 100 – (Dt/Du) × 100. Application of the pathogen and the biocontrol agents to the same site. In a set of experiments conducted under controlled conditions, suspensions of the pathogen and the biocontrol agent or agents were mixed before being applied to the detached strawberry leaflets. In the first experiments, the possible role of competition in biological control of Botrytis cinerea by Pichia guilermondii, Bacillus mycoides, or their mixture was examined. Live Vol. 92, No. 9, 2002
977
cells of the bacterium, the yeast, or their mixture, and conidia of Botrytis cinerea were supplemented with increasing concentrations of sugars, amino acids, and vitamins. In preliminary studies, several different proportions of the two biocontrol agents in the mixtures were examined. Because application of each of the biocontrol agents at a rate of 107 CFU/ml did not result in significantly different disease suppression compared with that achieved when each biocontrol agent was applied at a rate of 5 × 106 CFU/ml, the latter rates were used in the experiments. The sugars were glucose, fructose, and sucrose (at concentrations of 0.03 to 0.3%, wt/vol). The amino acids and vitamins were chosen from preliminary experiments in which four amino acids and four vitamins were tested. The amino acids and vitamins chosen for these experiments did not influence conidial germination of Botrytis cinerea or disease severity when applied at high concentrations. The amino acids were adenine (0.15 to 5.0 mg/ml) and histidine (0.08 to 2.5 mg/ml) and the vitamins were folic acid (0.015 to 0.5 mg/ml) and riboflavin (0.015 to 0.5 mg/ml). Conidial germination was evaluated in each of three different leaflet replicates. Disease severity was determined on five replicates (leaflets) per treatment, and the experiment was performed twice. Three factors were considered as possible contributors to Botrytis cinerea suppression by Pichia guilermondii and Bacillus mycoides: (i) the actual presence of live biocontrol cells; (ii) inhibitory compounds produced by the cells and attached to the cell surfaces; and (iii) inhibitory compounds secreted by the cells into the surrounding media. The compounds may or may not be sensi-
Fig. 1. Layout of the experiments conducted to evaluate the spatial effects of Pichia guilermondii, Bacillus mycoides, or their mixture on Botrytis cinerea. The treatments were as follows: A, control (Botrytis cinerea alone); B, together: the biocontrol agents and Botrytis cinerea were applied to the same site; C, main vein: the biocontrol agents were applied close to the main vein and Botrytis cinerea close to the margin of the leaflet; D, margin: the biocontrol agents were applied close to the margin of the leaflet and Botrytis cinerea close to the main vein; and E, parallel: the biocontrol agents and Botrytis cinerea were applied across parallel side veins. In treatments C to E, the biocontrol agents were applied at a distance of 0.5 cm from the drops containing Botrytis cinerea conidia. 978
PHYTOPATHOLOGY
tive to heat. In a set of experiments conducted on detached leaves, we attempted to estimate the partial contribution of each factor to Botrytis cinerea suppression. The biocontrol agents were grown for these experiments in LB liquid medium. In treatment one of the experiments, cells of the biocontrol agents were mixed with Botrytis cinerea before inoculation. In treatments two to four, cells of the biocontrol agents were separated from the medium by centrifugation (5,000 rpm for 10 min at 25°C). In treatment two, the separated cells were suspended in saline and mixed with Botrytis cinerea conidia. In treatment three, the cells were washed in saline twice before being mixed with Botrytis cinerea. In treatment four, the supernatant (possibly containing a compound or compounds secreted by the yeast or bacterial cells) was mixed with Botrytis cinerea conidia. Three- and tenfold concentrations of the supernatants were prepared by lyophilization. The experiments included four additional treatments, treatments five to eight, in which the liquid culture was boiled (100°C for 15 min) before preparing the cell suspensions. The heat killed the vegetative cells of Pichia guilermondii and Bacillus mycoides. Endospores of Bacillus mycoides may not be affected by heating and viable endospores could still exist. The heat-treated cells are referred to hereafter as heat-killed cells. Five drops of each of the treatments were placed on one side of the strawberry leaflets, and 5 drops containing only Botrytis cinerea conidial suspension (controls) were placed on the other side of the leaflets. There were five replicates (leaflets) per treatment. The experiment was performed twice. Application of the pathogen and the biocontrol agents on separate sites. In a set of experiments conducted on detached leaflets and with whole plants, the possibility of biocontrol agents affecting Botrytis cinerea development from a distance was tested. Spatial effects may result from secretion of an inhibitory compound (or compounds) that may be volatile or not, have acropetal or basipetal effects, or be dispersed by diffusion within the host tissues. Alternatively, effects on the pathogen may result from activating the plant defense systems. Spatial effects of the involved biocontrol agents, if active in vivo, may be governed by the position of the biocontrol agent relative to the position of the pathogen on the leaflet. The effect of the microorganisms’ position on the leaflets was examined in factorial experiments. The first factor in the experiments was the biocontrol agent, comprising three treatments: (i) Pichia guilermondii; (ii) Bacillus mycoides; and (iii) a mixture of Pichia guilermondii and Bacillus mycoides. The second factor in the experiment was the position of the biocontrol agents on the leaflets relative to the pathogen (Fig. 1), comprising five treatments: (i) control (Botrytis cinerea alone); (ii) together, the biocontrol agents and Botrytis cinerea were applied to the same site; (iii) main vein, the biocontrol agents were applied close to the main vein of the leaflet and Botrytis cinerea were applied close to the margin of the leaflet; (iv) margin, the biocontrol agents were applied close to the margin of the leaflet and Botrytis cinerea were applied close to the leaflet main vein; and (v) parallel, the biocontrol agents and Botrytis cinerea were applied near parallel side veins. Treatments one and two served as references for the potential development of the pathogen and the potential efficacy of the biocontrol agents, respectively. Treatments three and four were designed to identify whether the biocontrol agents induce acropetal or basipetal effects, respectively. Treatment five may help to identify whether the compound, or compounds, secreted by the biocontrol agents are dispersed within the leaf tissue irrespective of the direction of the xylem and phloem systems. Inoculation and incubation were done as previously described. Effects on conidial germination were determined on 50 Botrytis cinerea conidia in each of three different leaflet replicates. Effects on disease severity were determined on five replicates (leaflets) per treatment. The experiment was performed twice. To examine whether the biocontrol agents produce volatile compounds that influence germination of Botrytis cinerea conidia,
drops of cells of the biocontrol agents without medium, or supernatant prepared from liquid growth medium, were placed on glass slides. Cells were separated from the medium and the supernatants were prepared as previously described. Drops containing Botrytis cinerea conidia were placed on the same glass slides at a distance of 0.5 cm. The glass slides were placed in 13-cm-diameter petri plates and sealed with Parafilm. To maintain high relative humidity (>97%), pieces of wet cotton were placed inside the plates. Effects on Botrytis cinerea conidial germination were determined after 18 h of incubation at 20°C. There were three replicates (glass slides) per treatment, and 50 Botrytis cinerea conidia were inspected per replicate. The experiment was performed twice. The possibility that Pichia guilermondii or Bacillus mycoides induce resistance against Botrytis cinerea was examined on 10-month-old strawberry plants grown in 1-liter pots. The biocontrol agent, their mixture, or water was applied to the root zones of the plants (50-ml suspension at a concentration of 107 CFU/ml). Five days later, two fully expanded, mature leaflets on each plant were inoculated with the pathogen by applying 4 drops of Botrytis cinerea suspension per leaflet. The plants were covered with plastic bags to maintain high relative humidity for 24 h and placed in a growth chamber at 20°C. Disease severity was determined 7 days after inoculation. There were five replicates (plants) per treatment, and the experiment was performed twice. Scanning electron microscopy. Conidia of Botrytis cinerea and cells of the bacterium and the yeast were applied in 20-µl drops on strawberry leaves as previously described. Leaf samples (1 cm2) were taken for fixation after 6 and 24 h of incubation. The leaf samples were fixed in formalin/ethanol/acetic acid (4:25:1.5), dried in a critical-point drier, gold sputter-coated, and observed under a scanning electron microscope (JEOL T330; JEOL, Tokyo). Data analysis. Statistical analyses of the data were performed using the JMP-in software, version 3 for Windows (SAS Institute, Cary, NC). Analysis of variance was used to determine the influence of the treatments in the experiments on conidial germination and on disease severity of Botrytis cinerea. When the F values were significant, the least significant difference (LSD) values were calculated to identify significant differences between the treatments (P 0.05). For some analyses, multiple regression with dummy variables was employed. This is a general statistical procedure that encompasses analysis of variance and is useful in identifying quantitative effects of variables having distinct levels (26). The following regression equation was used:
