Biological Control Of Fire Blight Research Articles

Multisite Field Evaluation of Bacteriophages for Fire Blight Management: Incorporation of Ultraviolet Radiation Protectants and Impact on the Apple Flower Microbiome.

Fire blight, a disease of pome fruits caused by the bacterium Erwinia amylovora, has become increasingly difficult to manage after the emergence of streptomycin-resistant strains. Alternative antibiotics and copper are available; however, these chemicals have use restrictions in some countries and also can carry risks of phytotoxicity. Therefore, there is growing interest in biological-based management options, with bacteriophage (phages) showing promise, as these naturally occurring pathogens of bacteria are easy to isolate and grow. However, there are several technical challenges regarding the implementation of phage biocontrol in the field, as the viral molecules suffer from ultraviolet radiation (UVR) degradation and can die off rapidly in the absence of the host bacterium. In this work, we assessed the efficacy of Erwinia phages and a commercial phage product for blossom blight control in the field across multiple locations in the eastern United States. In these tests, disease control ranged from 0.0 to 82.7%, and addition of a UVR protectant only resulted in significantly increased disease control in 2 of 12 tests. We also analyzed microbial community population changes in response to phage application. Changes in bacterial community diversity metrics over time were not detected; however, relative abundances of target taxa were temporarily reduced after phage applications, indicating that these phage applications did not have deleterious effects on the flower microbiome. We have demonstrated that biological control of fire blight with phages is achievable, but a better understanding of phage-pathogen dynamics is required to optimize disease control efficacy.

Complete Genome Sequences of Five Gram-Negative Bacterial Strains Comprising Synthetic Bacterial Consortium "The Great Five" with Antagonistic Activity Against Plant-Pathogenic Pectobacterium spp. and Dickeya spp.

Complete Genome Sequences of Five Gram-Negative Bacterial Strains Comprising Synthetic Bacterial Consortium "The Great Five" with Antagonistic Activity Against Plant-Pathogenic Pectobacterium spp. and Dickeya spp.

Evaluation of a contact sterilant as a niche-clearing method to enhance the colonization of apple flowers and efficacy of Aureobasidium pullulans in the biological control of fire blight

Due to rapid expansion of organic apple production and issues with antibiotic resistance in conventional production, there is a need for novel strategies to protect orchards from outbreaks of fire blight caused by the bacterial pathogen Erwinia amylovora. The biological control material, Blossom Protect (consisting of two strains of the fungus Aureobasidium pullulans), has shown promising results in some apple growing regions but has shown reduced efficacy in humid apple-growing regions such as the Midwestern and Eastern United States. In an attempt to increase both flower colonization by A. pullulans and disease control efficacy, we applied a contact sterilant to clear the flower niche of native microbial populations 4 hr prior to application of Blossom Protect. Results from four years of field testing showed that application of the contact sterilant resulted in a significant reduction in populations of bacteria, fungi, and native yeast populations; however, all of these microbial populations recovered after 24 h. Examination of fire blight incidence revealed that use of the contact sterilant did not significantly impact disease control. Two applications of Blossom Protect, at 70–80% bloom and full bloom, exhibited the best efficacy in disease control over four years. In addition, we observed marginal differences in the marketability of the final fruit between treatments, each with less than 8% deemed unmarketable due to russet. Our results indicate that niche-clearing of apple flowers did not enhance flower colonization by A. pullulans, nor produce an increase in biological control efficacy. However, our current experimental results (2014–2018) do indicate consistency and an increase in efficacy of Blossom Protect compared to results from previous years.

NUTRITIONAL ENVIRONMENT INFLUENCES TRANSCRIPTION OF A PANTOCIN A BIOSYNTHESIS GENE IN PANTOEA VAGANS STRAIN C9-1

The histidine-reversible antibiotic pantocin A is produced by Pantoea vagans strain C9-1 and Pantoea agglomerans strains Eh252, EH318 and P10c and contributes significantly to efficacy of biological control of fire blight of pear and apple flowers caused by Erwinia amylovora. Antibiosis by pantocin A is observed on solidified defined media, such as MGA (MOPS gluconate medium amended with asparagine), but not on the solidified peptide-rich medium lysogeny broth (LB). It was not known if lack of antibiosis on LB was due to lack of pantocin A biosynthesis or to histidine in the medium reversing the effects of pantocin A. We used a green fluorescent protein (GFP) promoter probe in C9-1 to monitor transcription of the pantocin A biosynthesis gene paaA. GFP expression was higher in MGA than in LB. Addition of yeast extract or tryptone, components in LB, to MGA significantly decreased transcription of paaA. The observed lack of inhibition of E. amylovora by C9-1 in LB was positively correlated with paaA transcription. Addition of histidine or arginine, amino acids scarce on pome fruit stigmas, to MOPS gluconate medium (MG) decreased transcription of paaA compared to MG without amino acids. In contrast, asparagine and proline, amino acids that are abundant on pome fruit stigmas supported bacterial growth and transcription of paaA, which is in agreement with findings that pantocin A production is a key component of fire blight control by Pantoea spp. on floral stigmas.

Characterization of Myoviridae and Podoviridae family bacteriophages of Erwinia amylovora from Hungary - potential of application in biological control of fire blight

Twelve bacteriophage isolates of Erwinia amylovora, the causal agent of fire blight, were isolated from blighted apple, pear and quince trees from different sites in Hungary. According to morphological characteristics they were assigned to the order Caudovirales, two isolates belonging to the Podoviridae and ten to the Myoviridae families. Examining plaque morphology, host range and molecular characterization by PCR established that these phages are not identical neither to the three North American strains used as references nor the earlier isolated Hungarian Siphoviridae strains. Studying the efficacy of selected phages in apple blossoms and green pear fruit slices it was found that a combination of three phage isolates (ΦEaH2A, ΦEaH5K and ΦEaH7B) significantly reduced bacterial multiplication and fire blight symptoms as compared to untreated controls. Combined application of these new E. amylovora-specific phages as biocontrol agents may contribute to a better control of E. amylovora under field conditions.

Biological control of fire blight of apple and pear with antagonistic Lactobacillus plantarum

Lactic acid bacteria (LAB) can be a source of biological control agents (BCA) of fire blight disease. Several species of LAB are inhabitants of plants and are currently used as biopreservatives of food because of their antagonistic properties against bacteria, and are considered as generally safe. Candidates to BCA were selected from a large collection of LAB strains obtained from plant environments. Strains were first chosen based on the consistency of the suppressive effect against E. amylovora infections in detached plant organs (flowers, fruits and leaves). Lactobacillus plantarum strains PC40, PM411, TC54 and TC92 were effective against E. amylovora in most of the experiments performed. Besides, strains PM411, TC54 and TC92 had strong antagonistic activity against E. amylovora and also other target bacteria, and presented genes involved in plantaricin biosynthesis (plnJ, plnK, plnL, plnR and plnEF). The strains efficiently colonized pear and apple flowers; they maintained stable populations for at least 1 week under high RH conditions, and survived at low RH conditions. They were effective in preventing fire blight on pear flowers, fruits and leaves, as well as in whole plants and in a semi-field blossom assay. The present study confirms the potential of certain strains of L. plantarum to be used as active ingredient of microbial biopesticides for fire blight control that could be eventually extended to other plant bacterial diseases.

Mechanistically Compatible Mixtures of Bacterial Antagonists Improve Biological Control of Fire Blight of Pear

Mixtures of biological control agents can be superior to individual agents in suppressing plant disease, providing enhanced efficacy and reliability from field to field relative to single biocontrol strains. Nonetheless, the efficacy of combinations of Pseudomonas fluorescens A506, a commercial biological control agent for fire blight of pear, and Pantoea vagans strain C9-1 or Pantoea agglomerans strain Eh252 rarely exceeds that of individual strains. A506 suppresses growth of the pathogen on floral colonization and infection sites through preemptive exclusion. C9-1 and Eh252 produce peptide antibiotics that contribute to disease control. In culture, A506 produces an extracellular protease that degrades the peptide antibiotics of C9-1 and Eh252. We hypothesized that strain A506 diminishes the biological control activity of C9-1 and Eh252, thereby reducing the efficacy of biocontrol mixtures. This hypothesis was tested in five replicated field trials comparing biological control of fire blight using strain A506 and A506 aprX::Tn5, an extracellular protease-deficient mutant, as individuals and combined with C9-1 or Eh252. On average, mixtures containing A506 aprX::Tn5 were superior to those containing the wild-type strain, confirming that the extracellular protease of A506 diminished the biological control activity of C9-1 and Eh252 in situ. Mixtures of A506 aprX::Tn5 and C9-1 or Eh252 were superior to oxytetracycline or single biocontrol strains in suppressing fire blight of pear. These experiments demonstrate that certain biological control agents are mechanistically incompatible, in that one strain interferes with the mechanism by which a second strain suppresses plant disease. Mixtures composed of mechanistically compatible strains of biological control agents can suppress disease more effectively than individual biological control agents.

Genome Sequence of the Biocontrol Agent Pantoea vagans Strain C9-1

Pantoea vagans is a Gram-negative enterobacterial plant epiphyte of a broad range of plants. Here we report the 4.89-Mb genome sequence of P. vagans strain C9-1 (formerly Pantoea agglomerans), which is commercially registered for biological control of fire blight, a disease of pear and apple trees caused by Erwinia amylovora.

Field Evaluation of Biological Control of Fire Blight in the Eastern United States.

The bacterial antagonists Pseudomonas fluorescens A506, Pantoea agglomerans C9-1, and Pantoea agglomerans E325 and preparations of Bacillus subtilis QST 713 containing bacterial endospores and lipopeptide metabolites were evaluated for efficacy in controlling fire blight in Michigan, New York, and Virginia. When examined individually, the biological control materials were not consistently effective in reducing blossom infection. The average reduction in blossom infection observed in experiments conducted between 2001 and 2007 was variable and ranged from 9.1 to 36.1%, while control with streptomycin was consistent and ranged from 59.0 to 67.3%. Incidence of blossom colonization by the bacterial antagonists was inconsistent, and <60% of stigmata had the antagonists present in 12 of 25 experiments. Consistent control of blossom infection was observed when the biological control materials were integrated into programs with streptomycin, resulting in a reduction of the number of streptomycin applications needed to yield similar levels of control. Our results indicate that the prospects for biological control of fire blight in the eastern United States are currently not high due to the variability in efficacy of existing biological control options.

Implications of Pathogenesis by Erwinia amylovora on Rosaceous Stigmas to Biological Control of Fire Blight

As a prerequisite to infection of flowers, Erwinia amylovora grows epiphytically on stigmas, which provide a conducive habitat for bacterial growth. Stigmas also support growth of several other bacterial genera, which allows for biological control of fire blight; although, in practice, it is very difficult to exclude E. amylovora completely from this habitat. We investigated the dynamics of growth suppression of E. amylovora by comparing the ability of virulent and avirulent strains of E. amylovora to compete with each other on stigmas of pear, apple, and blackberry, and to compete with a co-inoculated mixture of effective bacterial antagonists. When strains were inoculated individually, virulent E. amylovora strain Ea153N attained the highest population size on stigmas, with population sizes that were approximately double those of an avirulent hrpL mutant of Ea153 or the bacterial antagonists. In competition experiments, growth of the avirulent derivative was suppressed by the antagonist mixture to a greater extent than the virulent strain. Unexpectedly, the virulent strain enhanced the population size of the antagonist mixture. Similarly, a small dose of virulent Ea153N added to inoculum of an avirulent hrpL mutant of Ea153 significantly increased the population size of the avirulent strain. A pathogenesis-gene reporter strain, Ea153 dspE::gfp, was applied to flowers and a subset of the population expressed the green fluorescent protein while growing epiphytically on stigmas of apple. These results are consistent with the hypothesis that virulent E. amylovora modifies the epiphytic habitat presented by the stigma through a pathogenesis-related process, which increases host resources available to itself and, coincidentally, to nonpathogenic competitors. Over nine orchard trials, avirulent Ea153 hrpL significantly suppressed the incidence of fire blight four times compared with six for the antagonist mixture. The degree of biological control achievable with an avirulent strain of E. amylovora likely is limited by its inability to utilize the stigmatic habitat to the same degree as a virulent strain.

EPIPHYTIC ERWINIAS TO CONTROL FIRE BLIGHT

Several epiphytic bacteria isolated from healthy apple and pear flowers were characterized by their phenotypic properties and by DNA sequence comparisons. Selected epiphytes from Australia and South Africa induced hypersensitive response (HR) on tobacco leaves and produced the exopolysaccharide levan. Sequence comparison of 16S rRNA and the housekeeping genes recA and gapDH showed a close relationship of these epiphytes to Erwinia amylovora and to other plant pathogenic bacteria of the genus Erwinia, however, the epiphytes form a separate group and are not pathogenic to apple and pear. Another close relative of E. amylovora, the epiphyte Erwinia billingiae did not produce levan and HR on tobacco, but it synthesized an acyl-homoserine lactone (AHL). The epiphytic strains from Australia and South Africa were antagonists against E. amylovora in assays on immature pear slices and on apple flowers. E. billingiae also showed significantly antagonism against E. amylovora. Using Real-Time PCR and plating samples on selective agar, we demonstrated that the epiphytic isolates provided effective growth suppression of the pathogen when apple flowers were co-inoculated with a high concentration of the epiphytes and low concentrations of E. amylovora. These epiphytes could possibly be used for biological control of fire blight.

Biological control of fire blight in pear orchards with a formulation of Pantoea agglomerans strain Eh 24

Biological control by using epiphytic bacteria against Erwinia amylovora has been considered as an alternative method for controlling the disease. Talc-based formulation of Pantoea agglomerans strain Eh-24 was applied at 30% and 100% bloom on two pear orchards which were selected from different locations in the Aegean Region in Turkey. Pear orchard trials were replicated for two years (1999 and 2000) in each place. Talc-based formulation of P. agglomerans strain Eh-24 was sprayed on pear trees which were naturally infected with E. amylovora. In the orchard trials conducted in 1999 and 2000, talc-based formulation of P. agglomerans strain Eh-24 reduced the percentage of blighted blossoms on pear orchards by 63% to 76%, approximately. Copper oxychloride+maneb was less effective in reducing the incidence of blossom infection by E. amylovora in each pear orchard than the bioformulation treatment. P. agglomerans strain Eh-24 labelled with StrR+ was applied at 30% and 100% bloom to monitor the colonization and population dynamics of P. agglomerans on pear blossoms. The population size of P. agglomerans strain Eh-24 strR+ on pear blossoms increased from 2x104 to 1.3x106 cfu per blossom over 18 days.

Antibiosis Contributes to Biological Control of Fire Blight by Pantoea agglomerans Strain Eh252 in Orchards

ABSTRACT Fire blight, caused by Erwinia amylovora, is the most serious bacterial disease of pear and apple trees. Biological control with strains of Pantoea agglomerans (syn. Erwinia herbicola) may provide an effective disease management strategy for fire blight. Most strains of P. agglomerans evaluated for suppression of fire blight produce compounds that inhibit the growth of E. amylovora in culture. The role of these inhibitory compounds in fire blight suppression in orchard environments has not been studied. In seven field trials in Oregon, we compared the population dynamics and disease suppression with P. agglomerans Eh252, a strain that produces a single antibiotic, with its near-isogenic antibiotic-deficient derivative, strain 10:12. Water or suspensions of Eh252 or 10:12 (1 x 10(8) CFU/ml) were applied at 30 and 70% bloom to pear or apple trees. Aqueous suspensions of freeze-dried cells of E. amylovora (3 x 10(5) CFU/ml) were applied at full bloom. Additional trees were treated with streptomycin or oxytetracycline at 30 and 70% bloom and in some experiments, 1 day after application of the pathogen. Population sizes of Eh252 or 10:12 on pear blossoms were estimated by spreading dilutions of blossom washes on culture media. Average population sizes of Eh252 and 10:12 on blossoms ranged from 10(5) to 10(7) CFU, and in five of six trials, the relative area under the population curve of Eh252 was not significantly different than that of its derivative 10:12. Both Eh252 and 10:12 reduced the growth of the pathogen on blossoms compared with inoculated water-treated controls. Eh252 significantly decreased the incidence of fire blight in six of seven field trials compared with the incidence on water-treated trees, and 10:12 similarly reduced the incidence of fire blight in four of seven trials. In three of seven field trials, trees treated with Eh252 had a significantly lower incidence of fire blight compared with trees treated 3 with 10:12. Overall,3 Eh252 reduced the incidence of fire blight by 55 +/- 8%, 10:12 by 30 +/- 6%, streptomycin by 75 +/- 4%, and oxytetracycline by 16 +/- 14%. The effectiveness of strain 10:12 compared with water treatment indicates that other mechanisms (e.g., competitive exclusion or habitat modification) also contribute to disease suppression by P. agglomerans. The increased suppression of fire blight by the parental strain Eh252 compared with the antibiotic-deficient mutant 10:12 indicates that antibiosis is an important mechanism of biological control of fire blight.

Field Performance of Antagonistic Bacteria Identified in a Novel Laboratory Assay for Biological Control of Fire Blight of Pear

Abstract A laboratory assay to identify antagonists of Erwinia amylovora on flowers was tested for its accuracy in predicting the performance of antagonists in controlling fire blight disease of pear trees. Bacterial strains isolated from pear flowers were evaluated in the laboratory for their ability to suppress growth of E. amylovora on flowers on cut branches. The growth of E. amylovora was assessed with a flower-freezing assay using a recombinant strain of the pathogen exhibiting an ice-nucleating phenotype. A number of antagonistic and nonantagonistic bacterial strains selected by this method were tested for their control of flower infection by E. amylovora on trees in the greenhouse and the field. There was a significant ( P ≤ 0.01) relationship between the ability of antagonists to suppress growth of E. amylovora in the flower-freezing assay and their ability to control fire blight on trees. Some of the strains identified conferred similar or better disease control than the biocontrol agent Pseudomonas fluorescens A506. None of the strains produced antibiotics inhibitory to E. amylovora on the culture media tested. All antagonistic and nonantagonistic strains survived well on pear flowers on trees and reached population sizes of approximately 10 5 to 10 6 cells/flower 3 days after their application. The use of this flower-freezing assay has promise in the rapid identification of biocontrol agents of fire blight.

Characterization of a Viral EPS-Depolymerase, a Potential Tool for Control of Fire Blight

ABSTRACT A 3.3-kb fragment of genomic DNA from bacteriophage Phi-Ea1h encoding an amylovoran-directed depolymerase lyase was sequenced, and three open reading frames (ORFs) were detected. The first ORF could encode a lysozyme and the second a holin that may form a pore supporting cell lysis by the lysozyme. The third ORF encodes a protein of 657 amino acids and deletion mutation in this DNA fragment abolished extracellular polysaccharide (EPS)-degrading activity. A putative promoter and a ribosome binding sequence were located in front of the gene. A polymerase chain reaction product spanning the gene was inserted into multi copy plasmids including fusions with a Histidine-tagged sequence to facilitate its purification on a nickel nitrilotriacetic acid column. Maximal activity of the purified protein was observed between pH 4 and 5 at 52 degrees C. Visualized by staining with fluorescein isothiocyanate-labeled lectin from Abrus precatorious, the enzyme degraded the EPS-capsules of Erwina amylovora. In virulence assays, no symptoms were detected for a low inoculum of an E. amylovora strain when pear slices were soaked in a solution of depolymerase in contrast to control slices without addition of the enzyme. Furthermore, gfp- or lux-labeled E. amylovora cells were not propagated, when their amylovoran capsules were removed by the depolymerase. The enzyme could be a tool for biological control of fire blight.

Effect of Nectar on Microbial Antagonists Evaluated for Use in Control of Fire Blight of Pome Fruits

ABSTRACT Under warm, dry conditions, Erwinia amylovora can become established in relatively high populations on apple (Malus domestica) or pear (Pyrus communis) flower stigmas, and subsequent wet conditions facilitate its movement to the flower hypanthium where infection generally is initiated through the nectarthodes. Research on biological control of fire blight has focused mainly on the flower stigma, and knowledge is lacking regarding the effect of nectar on microbial antagonists in the flower hypanthium. The biocontrol agents Pseudomonas fluorescens strain A506 and Pantoea agglomerans strain C9-1 were cultured in a basal liquid medium with various concentrations (0 to 50% total sugar) of sucrose or synthetic nectar (sucrose/glucose/fructose, 2:1:1). Strain A506 showed less growth and lower survival than strain C9-1 at high sugar levels, and A506 was less effective than C9-1 as a preemptive antagonist of E. amylovora in high-sugar media. Both antagonist strains were less tolerant to high sugar levels than E. amylovora (strain Ea153). The same bacteria were cultured in a medium with 25% total sugar consisting of various proportions of sucrose, glucose, and fructose, and growth response correlated strongly with solute potential. When 28 microbial strains were cultured in synthetic nectar (25% total sugar) and ranked based on growth, strains clustered according to taxonomic group. Yeasts were most osmotolerant, followed by strains of E. amylovora, Pantoea agglomerans, Bacillus spp., and Pseudomonas spp. Further studies done in planta are necessary to determine whether osmotolerance of antagonists is advantageous in the biological control of fire blight.

Inhibition of Monilinia oxycocci by bacteria isolated from a cranberry marsh

Bacteria inhibitory to Moniliniaoxycocci, the cranberry cotton ball pathogen, were identified. Eighty-three bacteria isolated from a cranberry marsh and Erwinia herbicola C9-1,which is being developed elsewhere for the biological control of fire blight of pome fruits, were tested fortheir ability to inhibit radial growth and conidial germination of M. oxycocci in vitro. IsolateBA35 from cranberry (tentatively identified as Pseudomonas syringae pv. morsprunorum) and E. herbicola C9-1 completely inhibited radial growth of M. oxycocci. BA35 and C9-1 were among the most effective of 21 bacteria tested for inhibition of conidial germination. Growth of bacteria in sclerotial extracts of M. oxycocci and Sclerotinia sclerotiorum was determined in order to select isolates that might be successful in degradingsclerotia, thereby reducing the viability of M.oxycocci. Populations of isolates 62 and S-18(identities unknown), and S-19 (tentatively identified as Micrococcus luteus), increased approximately3.0--5.0 log10 units in all extracts within 24 h. Populations of isolate S-10 increased by about 3.5log10 units in all sclerotial extracts within 48 h. Populations of isolate S-49 (identity unknown)were lower in extracts of M. oxycocci than S. sclerotiorum after 48-h. In buffer controls, bacterial populations remained stable or decreased over the 48-h period.

Crab Apple Blossoms as a Model for Research on Biological Control of Fire Blight

ABSTRACT Nonseasonal availability of pomaceous flowers could improve laboratory detection and prefield testing of biocontrol agents for fire blight of pear and apple. Crab apple was selected as a model because of its high flower productivity on 1-year-old wood, high susceptibility to fire blight, and availability from nurseries. Cultivars Manchurian and Snowdrift were manipulated to bloom once by transferring dormant nursery trees from a cold room to a greenhouse and a second time by defoliating trees and applying 1% cytokinin and 0.1% gibberellins to the buds with a brush. Different sets of trees were induced at different times to bloom, so that flowers were produced 12 months in the year. When known bacterial antagonists (Erwinia herbicola strain C9-1 and Pseudomonas fluorescens strain A506) were applied alone or in combination to the stigmas of detached crab apple blossoms prior to inoculation with the pathogen (E. amylovora strain Ea153), population interactions over time were comparable to those reported in previous studies involving pear or apple. In a subsequent series of experiments, the relative effects of 12 bacterial strains on stigmatic populations of strain Ea153 were similar for detached blossoms of crab apple in the laboratory, blossoms of intact crab apple trees in the greenhouse, and blossoms of pear and apple in the field. Additionally, when stigmas of detached crab apple blossoms were inoculated with antagonists (strains C9-1 and A506) and the pathogen, and later subjected to a 24-h wetting period, bacterial populations in the flower hypanthium increased and disease was suppressed. These studies indicate that crab apple blossoms can serve as a suitable model for year-round evaluation and study of biocontrol agents for fire blight.

A Method Involving Ice Nucleation for the Identification of Microorganisms Antagonistic toErwinia amylovoraon Pear Flowers

A method was developed for the rapid selection of antagonists capable of suppressing growth of Erwinia amylovora on pear flowers. The population size of E. amylovora on flowers pretreated with potential antagonists was estimated by a flower-freezing assay. E. amylovora harboring the iceC gene from Pseudomonas syringae on the stable plasmid pVSP61 expressed high levels of ice nucleation activity. The threshold freezing temperature of flowers colonized by the ice nucleation-active (Ice + ) strain of E. amylovora increased linearly with the logarithm of the population size of the pathogen. Bacterial and yeast strains were selected from the predominant microflora of pear flowers and tested for their ability to lower the threshold freezing temperature of flowers subsequently inoculated with the Ice + E. amylovora strain. For each antagonist, the proportion of 20 treated flowers that had frozen was evaluated when approximately 95% of the control (E. amylovora only) had frozen. Among 257 strains tested, 45 (4 yeasts and 41 bacteria) were capable of reducing freezing of inoculated flowers by 70% or more. The reduction in the proportion of pretreated flowers that froze relative to the control for a number of antagonists was correlated (R 2 = 0.61) with reduced E. amylovora population sizes. Over 50% of the antagonistic strains that reduced the proportion of frozen flowers by 50% or more during an initial screening conferred a similar reduction in the number of frozen flowers in a second test. Several strains were more antagonistic to E. amylovora in situ than P. fluorescens strain A506, which is used commercially for biological control of fire blight.

Effect of Antagonistic Bacteria on Establishment of Honey Bee-DispersedErwinia amylovorain Pear Blossoms and on Fire Blight Control

In field trials conducted in 1991 and 1992 at Medford, OR, and in 1992 at Wenatchee, WA, Pseudomonas fluorescens strain A506 and Erwinia herbicola strain C9-1 established epiphytic populations on pear blossoms and were effective antagonists for the biological control of fire blight. Both bacterial antagonists, water, or streptomycin sulfate were applied to trees at 30% and full bloom. Pear trees were challenged inoculated with freeze-dried cells of E. amylovora vectored to blossoms by honey bees. One week after full bloom, the antagonists were established in more than 95% of treated blossoms in Oregon in 1991 and Washington in 1992, but in less than 50% of blossoms in Oregon in 1992. At the same bloom stage, 41% (Oregon, 1991), 27% (Oregon, 1992), and 49% (Washington, 1992) of water-treated blossoms had detectable populations of E. amylovora, whereas trees treated with bacterial antagonists always had a significantly lower (p 0.05) percentage of blossoms with detectable E. amylovora populations: 18-20% (Oregon, 1991), 9-15% (Oregon, 1992), and 8-17% (Washington, 1992). In Oregon in 1991, only 4% of blossoms treated with bacterial antagonists supported populations of E. amylovora that exceeded 10(5) cfu per blossom compared with 19% of blossoms treated with water; however, suppression of population size of E. amylovora by bacterial antagonists was not apparent in 1992. In 1991, fire blight symptoms developed in 8, 0.1, and 1% of blossom clusters treated with water, streptomycin, or bacterial antagonists, respectively. In 1992, the percentage of diseased blossom clusters in these same treatments in Oregon averaged 44, 2, and 22%, respectively, and 9, 2.5, and 4%, respectively, in Washington