ERIC (Enterobacterial Repetitive Intergenic Consensus)-PCR was employed to generate stable and reproductive ERIC-PCR fingerprints of Ent. sakazakii ATCC51329. Moreover, this study also cloned and sequenced a major band of Ent. sakazakii (ATCC51329) ERIC-PCR fingerprints. The major band was amplified with primer ERIC2 and sequences extending primer ERIC 2 showed poor similarity with ERIC elements. A comparison of the nucleotide acid with other sequences available in the GenBank revealed 90% of identity with Ent. sakazakii ATCC BAA-894, and 73%-74% of identity with oligopeptiase gene or protease gene of some species from the Enterobacteriaceae family. Two primers were synthesized to develop and optimize an Enterobacter sakazakii-specific PCR based on regions of major band unique to Ent. sakazakii. The expected fragment was amplified from all of Ent. sakazkaii but not from the negative controls. As few as 10(2) CFU/ml of Ent. sakazakii of PCR were directly detected in the infant formulas. This was the case even in the presence of other bacteria. A comparison of traditional methods and new developed PCR in commercial foods suggested that without using API20-E test, the DFI chromogenic medium and FDA method showed 46.15% and 50% false positive respectively. Moreover, one false negative was observed with FDA method. In contrast, PCR was highly sensitive and specific to Ent. sakazakii. A high heterogeneity between Ent. sakazakii and the other microorganisms was found on expected fragment sequence. In addition, Ent. sakazakii ATCC51329 formed a separate branch with >5% divergence from the type strain ATCC BAA-894 and major strains.