Culture-independent characterisation of human faecal flora using rRNA-targeted hybridisation probes

Comparison of cultural and microscopic counts indicates that a significant fraction of human faecal flora is nonculturable. To characterise human faecal flora using quantitative dot-blot hybridisation, a set of ten ribosomal ribonucleic acid (rRNA)-targeted oligonucleotide probes were used to quantify rRNA abundance of different microbial groups within total RNA from faeces of ten individuals. The three domains and major groups of the domain bacteria were investigated. Cultural counts were compared with rRNA relative abundance for total Bacteroides, Bifidobacterium and Escherichia coli. Bacterial rRNA represented 78.2 +/- 4.3 % (range 59-94) of total rRNA. The archaeal and eucaryal rRNAs only showed a minor contribution to total rRNA (0.28 and 2.02 %, respectively), the added contribution of all three domains accounting for 80.5 +/- 4.5 % (range 61-100) of total rRNA. Additive probes targeting high G + C Gram-positives (HGC) and other bacteria (NHGC) hybridised with 1.4 and 80.6 % of bacterial rRNA, respectively. Total Bacteroides and enterics-group rRNAs accounted for 35.5 +/- 4.5 % and 2.7 +/- 0.7 % of bacterial rRNA, respectively. These results suggested a very significant contribution of low G + C Gram-positives (LGC) to total bacteria. Assuming that Bacteroides and enterics were the major Gram-negatives present, LGC could be calculated to account for 47 to 67 % of bacterial rRNA. Partial probing of LGC accounted for 24 +/- 5.5 % of bacterial rRNA. Cultural counts were more variable between individual faecal samples than were rRNA relative abundances for all three groups enumerated. Average counts of Bacteroides and Bifidobacterium, expressed as percent of total anaerobes, were two-fold higher than the corresponding rRNA indices, indicating a systematic overestimation of these groups. This could be due to an underestimation of total anaerobes. Our results indicate that culture-based techniques introduce a bias in the appreciation of the biodiversity of the human gut flora. LGC are apparently more susceptible to cultural underestimation. Their diversity as well as their functional contribution to the gut ecosystem may therefore need reconsideration. A complete molecular inventory of human faecal flora, currently underway in our laboratory, and probe-based technologies will certainly help to this end. (C) Inra/Elsevier, Paris