Molecular characterization of the human fecal flora using rRNA-targeted hybridization probes

Abstract

Culture-based anaerobic techniques of microbial identification and enumeration are central to studies of the human fecal flora. Nevertheless, they are time-consuming, require fresh samples and are not suited to large epidemiological investigations. Moreover, comparison of cultureand microscopic-counts indicate that a significant fraction of the human fecal flora is not yet culturable. Molecular techniques should allow more accurate characterizations of extensive numbers of frozen fecal samples. The ai___m_m of this study was to assess rRNA-based quantitative dot-blot hybridization for the characterization of Bacteroides, Bifidobacterium, and Escherichia. coli populations of the human fecal flora in comparison with cultural counts. Methods: Total anaerobes, Bacteroides, Bifidobacterium and E. coli populations were enumerated from fresh fecal samples obtained from 10 healthy volunteers. Total RNA was extracted from the same fecal samples after storage at -20°C; hybridized with specific 32p-labeled oligonucleotide probes, and quantified by radio-imaging. Results are expressed as percent of total viable anaerobes for culture-counts, and as percent of total bacterial rRNA for dot-blot quantification. They are given as mean ± SE. Results: CoUnts of total viable anaerobes were 7.9 109 cellsdg feces (wet weight). Total Bacteroides represented 59.9± 16.7 % of total viable anaerobes (3.4 109/g) and 35.5 ± 4.5 % of bacterial rRNA, Bifidobacterium represented 10.6 ± 3.3 % of total viable anaerobes (7.4 10S/g) and 4.1 ± 0.7 % of bacterial rRNA, and Escherichia coli 5.4 ± 3.5 % of total viable anaerobes (1.4 10S/g) and 2.7 ± 0.7 % of bacterial rRNA. Conclusion: Relative rRNA proportions for the different bacterial groups were in good aggreement with accepted levels based on cultural counts. The two-fold higher proportion of all bacterial groups analyzed by culture compared to quantitative dot-blot was likely due to a systematic underestimation of counts of total viable anaerobes. The present work demonstrates that the rRNA approach bears great potential for the analysis of the human fecal flora and circumvents the limitations of culture-based techniques.