The celA, manA, and celB genes from Caldocellulosiruptor saccharolyticus compose a cellulase-hemicellulase gene cluster and are arranged on a 12-kb C., saccharolyticus genomic fragment of the recombinant lambda bacteriophage NZP lambda 2. The beginning of a fourth open reading frame (celC) which was homologous to the C. saccharolyticus manA and celA genes was located at the 3' end of the 12-kb NZP lambda 2 genomic fragment. Genome-walking PCR was used to isolate DNA fragments downstream of the C. saccharolyticus celB gene, and the entire nucleotide sequence of celC was obtained. From the preliminary nucleotide sequence, celC appeared to encode yet another multidomain bifunctional enzyme (CelC) consisting of an N-terminal endo-1,4-beta-D-glucanase domain (75% similar to CelA domain 1), two central cellulose-binding domains, and a C-terminal endo-1,4-beta-D-mannanase domain (98% similar to ManA domain 1), However, upon completion of the celC sequencing, two -1 frameshifts were identified in the region encoding the putative CelC mannanase domain. The isolated CelC mannanase domain exhibited no beta-mannanase activity, which supported this observation. Recombinant PCR was used to correct the celC frameshifts by inserting the appropriate nucleotides into the gene. The repaired celC fragment containing the base insertions (manB) expressed strong beta-mannanase activity on soluble mannan substrates and showed significant activity on kraft pulp as judged by the release of reducing sugars.
