| | | The lack of suitable enzymes required for the breakdown of plant material represents a major bottleneck in the industrial scale roduction of cellulolytic biofuels. The cow rumen harbors microbes
that have evolved the molecular machinery to deconstruct polymeric lignocellulose efficiently into monomeric sugars. In this study, we mployed deep sequencing of the rumen microbiota associated tightly with switchgrass, a promising biofuel crop, to identify high-quality, full-length genes involved in the breakdown of the plant material. From paired end sequencing of a 200 bp insert library we generated 17 gigabases of 125 bp reads using the Illumina Genome Analyzer. To reduce computational complexity, de-novo assemblies were done on subsets of short sequence reads selected based on their homology to existing cellulose degrading enzymes. Sequences in this subset were extended stepwise through a greedy algorithm to extend partial contigs. Full length genes were identified by comparison to the set of known enzymes. In addition, enzymes that lack overall homology but harbor functional domains of interest were identified by an HMM-driven strategy. Combining the above two strategies almost 300 novel candidate enzymes were predicted. A large fraction of the predicted candidates could be cloned into expression vectors through PCR amplification of the initial genomic material from the rumen microbiota. For a subset of these genes we
validated the authenticity of their sequences by sequencing the amplified products. In order to assess the biochemical activity of the cellulolytic enzymes identified in the metagenomic data, functional
studies of the cloned genes, as well as a number of chemically synthesized genes, are underway.
In summary, from these studies we have dramatically increased the repository of lignocellulolytic enzymes exploiting a computational approach that can be generalized to assemble any family of genes and possibly whole microbial genomes from the deep sequencing of microbial communities. |
