A fingerprint database for characterization and identification of microbial contaminants

Effective process control and management should be able to trace spoilage organisms. Conventional identification methods of microbial contaminants are not reliable and specific enough for this purpose. The automated ribotyping system is a rapid molecular biological method fulfilling these requirements. However, its effective use is dependent upon comprehensive identification databases, representing strains from relevant sources. To augment the commercial database provided by the manufacturer, which now includes 4847 fingerprint patterns, VTT Biotechnology has created an own database, in which brewery contaminants and isolates from other food industry (seafood, poultry, meat), with especial reference to Listeria spp., have been our target groups. In addition to contaminants, useful strains like starter and probiotic cultures have also been ribotyped. Ribotyping was carried out using the automated ribotyping RiboPrinter® System (DuPont Qualicon, USA) with restriction enzymes EcoRI, PvuII or PstI. The system includes five stages: (i) DNA preparation and restriction, (ii) separation by gel electrophoresis directly linked to a membrane transfer, (iii) hybridization with a rRNA universal probe for detection, (iv) extraction and visualization of the pattern and (v) registration, characterization and identification of the patterns. The identification patterns of the Custom Identification Database (CID of VTT) were manually constructed from the generated ribogroups. Identification of strains for CID was confirmed by partial 16S rDNA sequencing. In order to form a dendrogram the relevant CIDs were transferred to the BioNumerics programme (Applied Maths, Sint-Martens-Latem, Belgium), where clustering was carried out using Pearson correlation and UPGMA (Unweight Pair Group Method using Arithmetric averages). The CID database includes isolates or strains of Lactobacillus spp., especially L. brevis and L. lindneri, Pediococcus spp., Pectinatus spp., Megasphaera cerevisiae and Obesumbacterium proteus, but also other lactic acid bacteria, enterobacteria, acetic acid bacteria, clostridia and pseudomonads have been ribotyped. Different restriction enzymes result in different fingerprints and discrimination. For example, 26 L. brevis strains have generated 14 different ribogroups (ribotypes) with EcoRI. For identification and characterization of Listeria monocytogenes there are 70 CIDs: 35 with EcoRI, 25 with PvuII and 10 with other enzymes. The most important property of ribotyping is the possibility to characterize strains below the species level, which has been successfully applied in tracing of contamination routes, in detection of troublesome "house types" and in evaluation of cleaning procedures in industrial processes. Considering the good discrimination power of ribotyping (even better than that of partial 16S rDNA sequencing), the database created will be extremely useful for reliable and rapid identification of bacteria isolated from food industry.