A STUDY OF DIVERSITY IN ERWINIA CAROTOVORA SUBSPECIES ATROSEPTICA IK TOTH, LJ HYMAN and PL TOTH Fungal and Bacterial Plant Pathology Dept., Scottish Crop Research Institute, lnvergowrie, Dundee. DD25DA,UK Background and objectives Erwinia carotovora subspecies atroseptica (Eca) is a bacterial plant pathogen of temperate regions responsible for blackleg disease of potato stems and soft rot of tubers in storage. To improve control measures against Eca, it is important to increase our understanding of the ecology of the pathogen and the epidemiology of blackleg and soft rot diseases. One aspect of this understanding is a study of the diversity of Eca in the environment. Until recently, studying the diversity of Eca, and the other major soft rot erwinias, E. c. carotoyora (Ecc) and E. chrysanthemi (Ech), was hampered by a lack of suitable methods. Eca, Ecc and Ech have traditionally been differentiated at species and subspecies level using biochemical and immunological characters. However, a number of typing methods, particularly those based on molecular analyses, have recently been used to study diversity within the soft rot erwinias, with varying degrees of success. In general, these studies show more diversity within Ecc and Ech than within Eca, and the diversity within Ech and, to a lesser extent, Ecc, has been correlated with host range and geographic location. Eca has previously been typed using PCR-RFLP, biochemical characters (API systems), fatty acid profiling, double immunodiffusion (DID), phage typing and RAPD analysis [1]. With the exception of RAPD analysis, these methods show Eca to be a highly homogeneous subspecies. The aim of this study was to develop molecular and physiological typing methods to increase the level of discrimination between Eca strains for unproved epidemiological analyses. Materials and methods Three methods; amplified fragment length polymorphisms (AFLP), a new phage typing system, and carbon utilisation analysis (Biolog system), were developed and tested on 60 Eca strains. In addition, RAPD analysis, being the most successful method previously used, and DID, because of its traditional role in Eca diversity studies, were also tested. Data produced by the above methods, with the exception of DID, were converted to binary form and dendrograms generated using the simple matching coefficient and the unweighted average pair group method. Results and conclusions DID was the least discriminatory method used, due to the predominance of serogroup 1 strains in the test group. In contrast, Biolog, phage typing, RAPD, and AFLP analyses, showed a high degree of variability between strains. A clear relationship was seen between serogroups and phage types, i.e. only serogroup I strains were phage-sensitive. This is not surprising, however, since all phages were originally isolated on a serogroup I Eca strain. Although this limits the usefulnness of the method at present, other Eca serogroups could be used to isolate non-serogroup I-specific phages. RAPD analysis produced nine sub-groups, although the relationship between these sub-groups and the nine produced in previous studies [1] is unclear, since different strains were used in each case. Due to the random nature of primer attachment, RAPD analysis was difficult to reproduce. Its ability to analyse DNA sequence variation over the whole chromosome, however, proved to be a useful method of differentiating strains. Using AFLPS, another method of random DNA analysis for comparison with RAPDS, up to 80 amplified bands were produced from each strain, with the number and size of these bands varying between strains. While strain groupings produced by AFLPs were similar to those produced by RAPDS, AFLPs were able to further differentiate strains within these groupings. Contrary to previous results, considerable diversity has now been demonstrated between in Eca strains. However, any relationships between the groupings produced by these methods and Eca ecological and virulence potential remains to be determined. Meanwhile, the demonstration of Eca diversity would allow the study of specific Eca populations in the environment, to further our understanding of blackleg aetiology and epidemiology. This work is currently ongoing in our laboratory. References 301-309. 1. Maki-Valkama T, Karjalainen R, 1994. Annals of Applied Biology 125,