We confirmed primer coverage and specificity in silico. The primer sequences (Matsuki et al., 2002) used in the present study matched with almost all rumen Prevotella sequences retrieved from the database and were specific for Prevotella, while the primers used by Stevenson & Weimer (2007) could anneal both Prevotella and Bacteroides. Therefore,

their primer set might have amplified ruminal Bacteroides, which have been frequently detected in previous analyses (Koike et al., 2003; Edwards et al., 2004), leading to overestimation of Prevotella. The RBB+C DNA extraction method that we used in ICG-001 nmr this study gives not only a high DNA yield, but it also produces superior results in PCR-based studies of diversity (Yu & Morrison, 2004), which is indicative of a more complete lysis and representation

of microbial community present in such samples. However, due to the animal species difference, it is likely that the relative abundance as well as the distribution of different Prevotella could be different in cattle learn more and sheep. Our phylogenetic analysis of Prevotella 16S rRNA gene sequences supports the findings of the quantification studies that indicated the predominance of uncultured strains. The majority (87.8%) of Prevotella clones had <97% sequence similarity to characterized rumen Prevotella,

which suggests that uncultured Prevotella are more abundant than cultured ones. Interestingly, the uncultured Prevotella clones were detected in similar proportions in both diets, suggesting their importance in ruminal fermentation of hay as well as concentrate diets. From the DGGE analysis, the common banding positions for both dietary conditions partially explain the versatile nature of Prevotella spp. reported previously (Avgustin et al., 1994, 1997; Matsui et al., 2000). In the phylogenetic tree, OTU37 and OTU51, which are composed of clones from both libraries, probably represent those rumen GBA3 Prevotella involved in the breakdown of both hay- and concentrate-based diets. However, findings from DGGE and clone library analyses suggested the existence of diet-specific members of Prevotella. DGGE profiles tended to cluster according to the diet given, and this result provided molecular evidence for the presence of diet-specific subpopulations of Prevotella that might be involved in the degradation of either a hay or a concentrate diet. The phylogenetic relationship of sequences of the libraries for each dietary condition supported the DGGE observation. libshuff comparison of the two libraries confirmed significant differences (P=0.