Cellulose in particular and wood in general are valuable biomaterials for humanity, and cellulose is now also in the spotlight as a starting material for the production of biofuel. Understanding the processes of wood formation and cellulose biosynthesis could therefore be rewarding, and genomics and proteomics approaches have been initiated to learn more about wood biology. For example, the genome of the tree Populus trichocarpa has been completed during 2006. A single-gene approach then has to follow, to elucidate specific patterns and enzymatic details.

This thesis depicts how a gene encoding a membrane-anchored cellulase was isolated from Populus tremula x tremuloides Mich, how the corresponding protein was expressed in heterologous hosts, purified and characterized by substrate analysis using different techniques. The in vivo function and modularity of the membrane-anchored cellulase was also addressed using overexpression and complementation analysis in Arabidopsis thaliana.

Among 9 genes found in the Populus EST database, encoding enzymes from glycosyl hydrolase family 9, two were expressed in the cambial tissue, and the membrane-anchored cellulase, PttCel9A1, was the most abundant transcript. PttCel9A1 was expressed in Pichia pastoris, and purified by affinity chromatography and ion exchange chromatography. The low yield of recombinant protein from shake flask experiments was improved by scaling up in the fermentor. PttCel9A1 was however highly heterogenous, both mannosylated and phosphorylated, which made the protein unsuitable for crystallization experiments and 3D X-ray structure determination. Instead, a homology model using a well-characterized, homologous bacterial enzyme was built. From the homology model, interesting point mutations in the active site cleft that would highlight the functional differences of the two proteins could be identified. The real-time cleavage patterns of cello-oligosaccharides by mutant bacterial enzymes, the wildtype bacterial enzyme and PttCel9A1 were studied by 1H NMR spectroscopy, and compared with results from HPAEC-PAD analysis. The inverting stereochemistry for the hydrolysis reaction of the membrane-anchored poplar cellulase was also determined by 1H NMR spectroscopy, and it was concluded that transglycosylation in vivo is not a possible scenario.

The preferred in vitro polymeric substrates for PttCel9A1 were shown to be long, low-substituted cellulose derivatives, and the endo-1,4--glucanase activity was not extended to branched or mixed linkage substrates to detectable levels. This result indicates an in vivo function in the hydrolysis of “amorphous” regions of cellulose, either during polymerization or crystallization of cellulose. In addition, overexpressing PttCel9A1 in A. thaliana, demonstrated a correlation with decreased crystallinity of cellulose. The significance of the different putative modules of PttCel9A1 was investigated by the construction of hybrid proteins, that were introduced into a knock-out mutant of A. thaliana, and the potential complementation of the phenotype was examined. A type B plant cellulase catalytic domain could not substitute for a type A plant cellulase catalytic domain, although localization and interaction motifs were added to the N- and C-terminus.