ORIGINAL ARTICLE Comprehensive detection of bacterial carbohydrate-active enzyme coding genes expressed in cow rumen Takumi SHINKAI, 1 Makoto MITSUMORI, 1,2 Ahmad SOFYAN, 2 Hiroyuki KANAMORI, 3 Harumi SASAKI, 3 Yuichi KATAYOSE 3 and Akio TAKENAKA 1,3 1 NARO Institute of Livestock and Grassland Science, 2 Graduate School of Life and Environmental Sciences, University of Tsukuba, and 3 National Institute of Agrobiological Sciences, Tsukuba, Ibaraki, Japan ABSTRACT To find the abundant and characteristic fibrolytic enzyme-coding gene expressed in fiber-associating microbiota, a metatranscriptomic data set was obtained from fiber-associating microbiota, and it was compared with that of rumen fluid- floating microbiota and two metagenomic data sets. Fibrolytic rumen bacteria associate with plant polysaccharide and hydrolyze it in the rumen. We obtained a metatranscriptomic assembly from fiber-associating microbiota in three ruminally fistulated Holstein cows fed timothy (Phleum pratense) hay. Each metatranscriptomic data set involved over a thousand of the glycoside hydrolase (GH) gene transcripts that accounted for about 1% of total protein coding gene transcripts. Three-quarters of the total GH gene transcripts were dominated by non-structural oligosaccharide-acting hydrolase gene transcripts. In the fiber-associating microbiota, endo-cellulase coding gene families, especially GHs 9 and 5, were abundantly detected, and GHs 9, 11, 30 and 43, carbohydrate esterase 8 and carbohydrate-binding module 6 were characteristically detected. Most fibrolytic gene transcripts assigned to Fibrobacter succinogenes were detected in fiber-associating sections, and GHs 45, 44, 74, 11, 30 and 16 were Fibrobacter-characteristically detected. The metatranscriptomic assembly highlighted the characteristic fibrolytic enzymes expressed in the fiber-associated rumen microbiota and offered access to the fibrolytic activities in each fibrolytic bacteria. Key words: cellulase, Fibrobacter succinogenes, metatranscriptomic analysis, plant polysaccharide, rumen. INTRODUCTION The lignocellulosic biomass is one of the most abundant and sustainable resources on earth. However, its decon- struction is an obstacle in global carbon cycling because of its complexity (Lynd et al. 1991). Ruminants retain microbiota that are able to effectively digest the forage in their rumen, where bacterial enzymes play an important role in the deconstruction of plant polysaccharides (Flint et al . 2008). Thus, the ruminal fibrolytic system has been a focus of research seeking to identify lignocellulosic enzymes for various industries, including biofuel produc- tion (Lynd et al . 2002; Dodd & Cann 2009; Hess et al. 2011; Kuhad et al. 2011). Rumen bacteria form a fibrolytic consortium on plant materials to develop interspecies partnerships for nutrient processing (Koike et al. 2003; Brulc et al. 2009). A small subunit (SSU) ribosomal DNA (rDNA) sequence-based phylogenetic analysis revealed that the rumen fibrolytic consortium consists of a huge variety of bacterial species and a large bacterial population yet to be cultured (Kobayashi et al. 2008). It is generally accepted that Fibrobacter, Ruminococcus and Clostridium have major roles in ruminal plant polysaccharide degradation (Koike & Kobayashi, 2001; Miron et al. 2001; Krause et al. 2003; Shinkai & Kobayashi 2007). Both Fibrobacter and the Ruminococcus-Clostridium group associate with plant ma- terials to compete with other bacteria for the use of the solubilized plant polysaccharide (Miron et al. 2001), that is bacterial cell with fibrolytic activities should associate with the plant fiber. This characteristic is useful in attempts to isolate the fiber-associating fibrolytic consor- tium from the complex rumen microbiome (Brulc et al. 2009; Hess et al. 2011). Genomics and metagenomic research has contributed not only to our understanding of the plant polysaccha- ride degradation system, but also to the improvement of the gene database (Brulc et al. 2009; Hess et al. 2011; Suen et al. 2011). However, only limited information is available for estimating bacterial functional activities Correspondence: Takumi Shinkai, NARO Institute of Livestock and Grassland Science, Tsukuba, Ibaraki 305–0901, Japan. (Email: tshinkai@affrc.go.jp) Received 16September 2015; accepted for publication 17 November 2015. © 2016 Japanese Society of Animal Science doi:10.1111/asj.12585 Animal Science Journal (2016) ••, ••–••