This study was conducted to develop a rapid numerical procedure for the analysis of gas chromatograms of lactic acid bacteria fatty acid methylesters (FAMEs), along with an examination of some experimental conditions which affect the fatty acid composition of these bacteria. FAME determination was carried out using high resolution gas chromatography (HRGC). Although the nature and the proportion of fatty acids differed among strains, some major components, n-C14:0, n-C16:1, n-C16:0, n-C18:1, n-C18:0, and C19 cycl, were found as a group to represent more than 90% of the whole cellular fatty acids. The differences found in the relative composition of the long-chain fatty acids in paired chromatograms were used to calculate a “distance coefficient”, based on the differences found for the most important fatty acids, after their prior ranking. The procedure was initially validated with identified species, used as reference strains; then, unknown lactic acid bacteria isolates were compared to these references. The method proved to be useful for rapid comparisons between strains, provided a strict standardization occcurred prior to routine application.

The cellular fatty acid compositions of 29 strains ofYersinia pestis representing the global diversity of this species have been analyzed by gas-liquid chromatography to investigate the extent of fatty acid polymorphism in this microorganism. After culture standardization, all Y. pestis strains studied displayed some major fatty acids, namely, the 12:0, 14:0, 3-OH-14:0, 16:0, 16:1ω9cis, 17:0-cyc, and 18:1ω9trans compounds. The fatty acid composition of the various isolates studied was extremely homogeneous (average Bousfield's coefficient, 0.94) and the subtle variations observed did not correlate with epidemiological and genetic characteristics of the strains. Y. pestis major fatty acid compounds were analogous to those found in other Yersiniaspecies. However, when the ratios for the 12:0/16:0 and 14:0/16:0 fatty acids were plotted together, the genus Yersinia could be separated into three clusters corresponding to (i) nonpathogenic strains and species of Yersinia, (ii) pathogenicYersinia enterocolitica isolates, and (iii) Yersinia pseudotuberculosis and Y. pestis strains. The grouping of the two latter species into the same cluster was also demonstrated by their high Bousfield's coefficients (average, 0.89). Therefore, our results indicate that the fatty acid composition ofY. pestis is highly homogeneous and very close to that ofY. pseudotuberculosis.

The cellular fatty acid composition of about 160 strains of Bacillus, lactic acid bacteria, Enterobacteriaceae, and Staphylococcus was analyzed in order to identify foodborne bacteria. In a first step, the reproducibility of the procedure, from the bacterial growth to the extraction and analysis conditions, was evaluated. The stability of fatty acid composition under controlled conditions was confirmed through high similarity levels (So ≥ 94). Secondly, intraspecies similarity was estimated. The existence of low values (So ≤ 79) was demonstrated within each of the four groups of strains investigated, which stressed the importance of analyzing a high number of reference strains in each species. The intraspecies similarity was maximal under optimal growth conditions. Finally, interspecies similarity was compared to intraspecies similarity. The best results of discrimination between species was obtained for the heterogeneous genus Bacillus, for which the 12 species studied could be separated into 6 clusters at a similarity level of 79. The association of 27 species of lactic acid bacteria into 8 clusters should be useful to confirm results of other characterization methods. Only 3 clusters were formed for the 16 Enterobacteriaceae species studied. The genus Salmonella could not be discriminated. The 10 Staphylococcus species were separated into 5 clusters. Coagulase-positive and -negative strains were not differentiated. Since the fatty acid composition did not usually appear specific to single species, characterization of foodborne bacteria by the analysis of their cellular fatty acids should be used to complement other taxonomic methods.