2.10.2S
BACTERIAL INTERACTIONS IN THE PHYLLOSPHERE AS A BASIS FOR BIOLOGICAL CONTROL OF BACTERIAL DISEASES

S LINDOW

Department of Plant and Microbial Biology, University of California, Berkeley, USA

Plant pathogenic and ice nucleation active bacteria commonly occur on the surface of healthy plants as epiphytes where, under the correct environmental conditions, they can result in disease or frost injury, respectively. The likelihood of disease or frost injury is proportional to the logarithm of the size of these epiphytic populations. Therefore, biological control can be successful if antagonistic microbes can achieve a reduction in the epiphytic populations of these pests. Pre-emptive competitive exclusion is an important mechanism of interaction among bacteria on leaves. Antibiosis does not appear to be a major factor determining interactions of bacteria on leaf surfaces, unlike in the rhizosphere. The population size of most epiphytic bacteria are limited by the same resource; on plants of normal nutrition carbon sources appear to be the most limiting resource since addition of C sources but not N or P sources increases population sizes of epiphytes. Bacteria can exhibit both balanced and directed competition for the limiting resources on plants. The effectiveness of antagonists as biological control agents is not correlated with the degree to which they exhibit directed competition, but instead is correlated with the degree to which their niche overlaps with that of the target pathogen on leaf surfaces. The effectiveness of biological control agents such as Pseudomonas fluorescens strain A506 is due to their wide nutrient utilization profile, enabling them to deplete limiting resources on plants when inoculated in advance of pathogens. That is, such strains exhibit a high degree of niche overlap with that of the pathogen which must compete with such strains for limiting resources. Competition on leaves may be considered a local phenomenon. Bacteria are highly aggregated on the surface of leaves, and a majority of the cells are located in a few isolated sites that are not closely associated with any morphological feature of the leaf. A significant amount of carbon-containing nutrients remain on leaves even after bacterial population sizes have increased to high and stable levels. Despite this, the use of recombinant bacteria harbouring reporter gene fusions that enable the bacteria to produce a unique phenotype that is dependent on the presence of environmental signals, such as sugars, have been used as 'biosensors' for resources such as sucrose and have revealed that nutrient levels decrease dramatically in the vicinity of bacteria as growth ceases on leaves. Nutrients are therefore probably depleted locally, but remain in locations on leaves where, because of other limiting factors, bacterial colonization is not favoured. The populations of bacteria on some plants such as citrus are driven primarily by immigration and not by growth. On such plants bacterial populations are relatively low compared with other nearby plants and population sizes increase only slowly with time and are influenced strongly by the proximity of the plant to plant species nearby that harbor large epiphytic bacterial populations. In such cases biological control by competitive exclusion would not be expected to operate since pathogens or ice nucleation active bacteria would derive few resources on leaves for which competition could occur. In contrast, biological control has been most effective on plants such as pear where newly opened flowers provide a rich nutritional resource that otherwise is rapidly exploited by pathogens or ice nucleation active bacteria in the absence of antagonistic bacteria.